Redefining the Typology of Land Use
in the Age of Big Data
MASSACHU§TTIltfE
OF TECHN9LOGY
by
JUN 19 2014
Liqun Chen
UB RARIES
B. Eng., City Planning
Peking University, 2012
SUBMITTED TO THE DEPARTMENT OF URBAN STUDIES AND PLANNING
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER IN CITY PLANNING
AT THE
OF TECHNOLOGY
INSTITUTE
MASSACHUSETTS
JUNE 2014
©2014 Liqun Chen. All rights reserved.
The author hereby grants to MIT permission to reproduce
and to distribute publicly paper and electronic
copies of this thesis document in whole or in part
in any medium now known or hereafter created.
Signature redacted
Signature of Autho r:
Department of Urban Studies and Planning
May 23, 2014
Certified by:
_
Signature redacted
d0
Jinhua Zhao
Assistnt Pgfesn of Urban Planning
Signature redacted Thesis Supervisor
Accepted by:
Associate Prifess
Christopher Zegras
C air, MCP Committee
Department of Urban Studies and Planning
2
Redefining the Typology of Land Use
in the Age of Big Data
by
Liqun Chen
Submitted to the Department of Urban Studies and Planning
On May 23, 2014 in Partial fulfillment of the
Requirements for the Degree of
Master in City Planning
ABSTRACT
Land use classification is important as a standard for land use description and management.
However, current land use classification systems are problematic. Labels such as
"residential use" and "commercial use" do not fully reveal how the land use is used in
terms of function, mix use and changes over time. As a result, land use planning is often a
natural prompt of segregation; Land use is poorly connected with other fields of urban
studies such as transportation and energy consumption. The problems of land use are partly
because land use classification has been an expediency rather than of rigorous thought.
However, recent researches about land use classification have mainly focused on the
methods of estimating land use types, without challenging the conventional instructional
definition of land use typology itself. In contrast, this thesis aims to ask a more fundamental
question: what are the elements, the principles, and the process to build the land use
typology for given purposes.
This thesis accordingly proposes the syntax of developing a land use typology, where five
basic elements compose the framework of land use description: land use function, land use
intensity, land use connectivity, probability and scale. While the elements are abstract
concepts, when developing a land use typology, each of them could be defined with
specific measures for purposes such as land use planning, land use management, energy
analysis, transportation study. After the land use typology is composed with the defined
3
elements, it can be applied to examine land mixed use, land use conflict, land use change
and estimation.
The syntax then proposes the basic principles and process to develop a satisfied land use
typology, with respect to the reliability and validity, the significance and necessity, the
measurability and operability, and the adaptability and flexibility. With that, this thesis
argues that beyond the theoretical definition, the practical context, such as data availability
or planning schema will influence the feasibility of a land use typology.
While the scope of the syntax could be limited by practical tools and availability of data,
the coming age of big data provides a changing context of land use typology. The followed
case study illustrates such a process of developing land use typology with geosocial
network data. The case develops a social media based land use typology, collects data for
two example cities: Boston, U.S and Shenzhen, China, and applies the defined land use
typology to classify their uses of land. As a result, Boston's land use I classified by its
function, intensity and the level of mix use; Shenzhen land use is classified by its intensity,
connectivity and the level of mix use.
Compared with the conventional land use classification systems, the social media based
typology provides a more comprehensive description of land use, with its focuses on
human activities of the city and multiple dimensions of urban land use. It also has
advantages with the flexibility and efficiency of data collection.
In conclusion, the syntax of land use typology highlights the process of building land use
typology, by defining the basic components of land use typology. It enables many
possibilities of land use description with the help of big data, and reserves enough space to
go beyond the existing tools and techniques. At last, the thesis proposes for future studies
on the different interpretations of the syntax, its application on planning tools and systems,
and potential for new types of land use.
4
Thesis Supervisor: Jinhua Zhao
The Edward H. and Joyce Linde Career Development Assistant Professor of Urban
Planning, Department of Urban Studies and Planning
Thesis Reader: Ying Long
Associate Professor, Beijing Municipal Institute of City Planning & Design
5
ACKNOWLEDGEMENT
Wrapping up this thesis, I finally notice this will be the end for now and a starting point
of another episode. It was difficult to imagine such a moment to leave MIT, where I have
struggled, enjoyed and appreciated.
Here I would like to thank all these people for their advice and support:
To Jinhua, for your passion, your guidance, the tireless review of my work and all of the
great ideas. I will never be able to go this far without your help. You have set me an
example of getting academic achievement with happiness. I could not have asked for a
better advisor.
To DUSP community members, for all of the inspiration, fun and support. From you, I
learned what a vibrant and diverse group would look like, and I am proud of being part of
it.
To my parents, for all the love, support, and understanding. You have been giving me the
courage to pursue what is right and important for me. Knowing that you will always be
there, I become more and more realized on where I come from and where I am going.
To my dear friends, for your friendship; for growing up together with me; for loving me
as who I am in the past 10 years; for the fun along the sea, river and lake; for all of the
memory at MIT.
6
TABLE OF CONTENTS
I
IN TRODUCTION .............................................................................................
13
1.1
The Problem s of Land U se Classification........................................................
13
1.2
Objectives / Research Questions .....................................................................
14
1.3
Fram ework of study .........................................................................................
15
1.4
U se of Term inologies.......................................................................................
16
A LAND USE AND PRACTICE ............................................................
U RBAN
18
2.1
U . S Land U se Classifications.........................................................................
18
2.2
China Land U se Classifications ....................................................................
24
2.3
Land U se Inform ation System s .......................................................................
26
2.4
Land Use Analysis ........................................................................................
28
2.5
Land U se Change and Estim ation ..................................................................
31
2.6
Brief Sum m ary ................................................................................................
31
THE SYNTAX OF LAND USE TYPOLOGY .................................................
34
3.1
Understanding Land Use................................................................................
34
3.2
The Elem ents of Land U se Typology ...........................................................
35
2
3 1
3.1.1
Function ..................................................................................................
35
3.1.2
Intensity..................................................................................................
36
3.1.3
Connectivity...........................................................................................
37
3.1.4
Probability ................................................................................................
37
3.1.5
Scale of Tim e and Space.........................................................................
40
3.3
Land U se Typology: a Package of the Elem ents.............................................
43
3.3.1
Packages for Energy Analysis ................................................................
44
3.3.2
Packages for City Planning.....................................................................
45
3.3.3
Packages for Land Resources M anagem ent ............................................
46
3.3.4
Packages for Transportation Study .........................................................
46
Generic Applications.......................................................................................
47
3.4.1
Land M ixed Use.......................................................................................
47
3.4.2
Land Use Conflict..................................................................................
48
3.4
7
3.4.3
3.5
Land U se Change and Future...................................................................
50
The Process of Application .............................................................................
51
3.5.1
The Principles ..........................................................................................
51
3.5.2
The Unit of M easure ................................................................................
52
3.5.3
The Steps..................................................................................................
55
4
BIG DATA AN D THE OPPORTUN ITIES .......................................................
57
4.1
The Concept of Big Data................................................................................
57
4.2
Opportunities for Urban and Land Use Studies ..............................................
59
4.2.1
Broader Resolution and Scale..................................................................
59
4.2.2
Dynam ics Data Collection and Analysis ................................................
60
4.2.3
Exploring the Com plexity of Cities .........................................................
61
4.2.4
Urban Prediction ......................................................................................
62
4.3
5 I
Land U se: Data and Beyond Data ..................................................................
62
CASE STUDY: TESTING THE SNTAX WITH BIG DATA..........................
65
5.1
Exam ple Cities ...............................................................................................
65
5.2
Big Data M ining and Process.........................................................................
67
5.2.1
M ain Datasets...........................................................................................
67
5.2.2
The Techniques to Acquire Data ...........................................................
69
Com posing the Typology ................................................................................
71
5.3.1
Topic of Interest/Purpose .........................................................................
71
5.3.2
Defined Scale and Unit of M easure .........................................................
71
5.3.3
D efined Elem ents....................................................................................
71
BO STON Result.............................................................................................
72
5.3
5.4
5.4.1
Data Briefing...........................................................................................
72
5.4.2
Unit of M easure ......................................................................................
75
5.4.3
Land U se Elem ents ..................................................................................
75
5.4.4
Generic Applications ...............................................................................
80
5.5.1
Types of Land U se ..................................................................................
85
5.5
5.5.1
SHEN ZHEN Result............................................................................................
Data Briefing ...........................................................................................
8
87
87
5.5.2
Unit of M easure ......................................................................................
89
5.5.3
Land U se Elem ents ..................................................................................
89
5.5.4
Land Use Application .............................................................................
94
5.5.5
Types of Land U se ..................................................................................
97
Reflections..........................................................................................................
99
5.6.1
Lim itations ...............................................................................................
99
5.6.2
Evaluation .................................................................................................
100
5.6.3
Im plem entation and Im pact ......................................................................
102
5.6
6
I
CON CLU SION ....................................................................................................
103
6.1
Value ................................................................................................................
105
6.2
Lim itations of the Study...................................................................................
105
6.3
Future Researches ............................................................................................
106
BIBLIOGRAPHY .............................................................................
108
A PPEN DIX .....................................................................................................................
112
A.
American Planning Association Land-Based Classification Standards............... 112
B.
China National Standard of Land Classification Gb/T 21010-2007....................
C.
China Code for Classification of Urban Land Use and Planning Standards of
Development Land GB 50137-2011 .......................................................
9
126
134
LIST OF FIGURES
Figure 1-1 the Framework of Study..............................................................................
15
Figure 1-2 the Relationship of the Terms Used In This Thesis .....................................
17
Figure 2-1 Land Use Map by Bartholomew ..................................................................
20
Figure 2-2 Zoning Map of New York City ..................................................................
Figure 2-3 APA Land Based Classification Standards ................................................
21
24
Figure 2-4 the Interaction of Land Use and Transportation: Impact of Traffic on
A ctiv ities...........................................................................................................................
Figure 3-1 Land Use as a Physical Dimension of Human Activities ............................
Figure 3-2 the Elements of Land Use ...........................................................................
30
34
Figure 3-3 the Probability Field of Land Use ................................................................
38
Figure
Figure
Figure
Figure
Figure
40
41
42
43
47
3-5
3-6
3-7
3-8
3-9
Activities with the Continuity of Time and Space .......................................
the Movement of Individual in Time and Space..........................................
Scales of Space and Time ...........................................................................
Land Use Typology as a Package of the Elements .....................................
the Various Focus of the Generic Applications .........................................
35
Figure 3-10 the Identification of Land Use Conflicts...................................................
Figure 3-11 the Prototype of Land Use With Probability Field.....................................
49
50
Figure 3-12 Raster Unit and a Raster Based Heat Map ................................................
52
Figure 3-13 Thiessen Polygon And an Example Map ..................................................
53
Figure 3-14 Building Footprint as an Object Based Unit System ................................
Figure 3-15 the Steps of Land Use Typology Development .........................................
Figure 4-1 The World's Capacity Change to Store Information ...................................
54
55
58
Figure 4-2 Scale And Resolution of Big Data ..............................................................
Figure 4-3 Live Singapore, City Decisions in Sync ....................................................
59
61
Figure 4-4 Urban Complexity Study: Population Morphology and the Road Network of
L ondon ..............................................................................................................................
61
Figure 5-1 Grid for Data Collection..............................................................................
70
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
73
74
76
77
79
80
81
82
82
83
5-2 Boston, the Output of Data.........................................................................
5-3 Boston, Parcels, Buildings and Roads..........................................................
5-4 Boston, Categories of Function..................................................................
5-5 Boston, the Count of Parcels by Function Types.......................................
5-6 Boston, the Intensity of Land Use................................................................
5-7 Boston, Mix Use of the City ......................................................................
5-8 Boston, Single Used Parcels by Function ..................................................
5-9 Boston, Dominated Used Parcels by the Dominated Function of Use ...........
5-10 Boston, Mix Used Parcels by the Most Frequent Type of Function..........
5-11 Boston, Land Use Diversity as an Index of Mix Use.................................
10
Figure 5-12 Boston, Land Use Change in a Week........................................................
Figure 5-13 Boston, Land Use Types ...........................................................................
84
85
Figure 5-14 Boston, Existing Land Use Typology .......................................................
86
Figure
Figure
Figure
Figure
Figure
Figure
Figure
5-15
5-16
5-17
5-18
5-19
5-20
5-21
Shenzhen,
Shenzhen,
Shenzhen,
Shenzhen,
Shenzhen,
Shenzhen,
Shenzhen,
the Output of Data ....................................................................
Categories of Function............................................................
the Count of Cells by Function Types .....................................
the Intensity of Land Use..........................................................
the Connectivity of Land Use...................................................
Single Use (a), Dominate Use (b) and Mix Use (c) of Land .......
Land Use Intensity And Mix Use ............................................
Figure 5-22 Shenzhen, Land Use Types ......................................................................
88
90
91
92
93
95
96
97
Figure 5-23 China, Existing Land Use Categories .......................................................
98
Figure 6-1 the Syntax of Land Use Typology ................................................................
103
11
LIST OF TABLES
Table 2-1 Urban Land Use Classification for Zoning ..................................................
19
Table 2-2 Functional Uses of urban land, Harland Bartholomew .................................
19
Table 2-3 Multiple Dimensional Land Use Classification ...........................................
23
Table 2-4 China National Standard of Land Classification .........................................
25
Table 3-1 an Interpretation of the Development for Energy Based Land Use Typology. 44
Table 3-2 an Interpretation of the Land Use Typology Elements in Different Scale of City
P lann in g ............................................................................................................................
45
Table 3-3 an Interpretation of the Development for Land Resource Based Land Use
T ypology ...........................................................................................................................
46
Table 3-4 an Interpretation of the Development for Transportation Based Land Use
46
Typo logy ...........................................................................................................................
Table 5-1 U. S Land Cover by Type (in millions of acres) ...........................................
65
Table 5-2 China Land Cover by Type (in ten thousands hectares)............................... 66
Table 5-3 Quick Fact of the Two Example Cities .......................................................
67
Table 5-4 Boston, the Input of Query ...........................................................................
72
Table 5-5 Boston, the Output of Query ........................................................................
72
Table 5-6 Boston, Building Footprint Data Fields........................................................
75
Table 5-7 Shenzhen, the Query of Data.........................................................................
87
Table 5-8 Shenzhen, the Output of Data.......................................................................
87
Table 5-9 Evaluation of the Social Media Based Land Use Typology........................... 100
12
1
INTRODUCTION
1.1
The Problems of Land Use Classification
Land use is the use of land by human beings. As Albert Guttenberg, a seminal scholar on
land use has emphasized, it is important for planners to use a common language. Land
use classification, in his opinion, is one of the important terms that planners should share
in the common language.
The terms of "land use classification" and "land use planning" remain vague, "too many
planners use their language unreflectively(A. Z. Guttenberg, 1993). There has been
plenty of literature, pointing out the problems of current land use classification system
and calling for reform of zoning/land use regulation. The problems are mainly on several
aspects:
" Function and urban form. When talking about land use, it is difficult to tell the
actual development patterns on the ground, about how the land is being used. That
is because land use is labeled simply by categories such as "residential" or
"commercial". In the past, this classification and its relating administrative system
have been used as one of the main methods to describe and regulating urban land
use. However, this system could no longer fulfill the need of understanding urban
land use.
" Mix use and diversity. Mix use is not only happening naturally in the city, but
also encouraged by planning ideas for the sake of health and life qualify. Current
land use classification systems are not able to represent the degree and quality of
mixed use.
* Urban changes. As Blumenfeld pointed out as early as 1967, "the static concept of
the city is no longer valid. It is constantly changing and growing," which becomes
extremely true in our age. With increasing changes in the urban environment,
particularly in rapidly developing areas and extended metropolitan regions, the
process of updating land use, and with greater frequency, has requested significant
13
amounts of labor and time. As a consequence, retrieving timely accurate land use
information so as to maintain the pace of urban development is a critical challenge
confronting urban planners. Even regardless of whether urban population and land
use change are slow or rapid, it is difficult to measure actual changes in urban land
use because the definition of land use is confused if not absent.
As the result of using these systems, land use planning is often a natural prompt of urban
spatial segregation. Moreover, current land use classification systems have a poor
connection with other fields of urban studies. For example, in transportation studies, land
use is recognized as one of the key factors to influence behavior patterns and travel
demand. However, these studies could not rely much on existing land use data because
the classes of land use could not tell much on daily activities.
1.2
Objectives / Research Questions
With the problems of current land use classification systems, the research started with the
question of what should be a valid land use typology, or possible typologies that are
beyond the conventional categories of land use. Soon I found that it is almost impossible
to develop such a system that is generic enough for multiple purposes and at the same
time fit into a specific practical context.
Then the research question turned into what should be the components, the principles,
and process to build a valid land use typology. The idea is simple, with the more and
more complex and diverse need of land use related studies, there could be some key
elements of land use description. If we can identify the elements and rules of composing
a land use typology, then planners and researchers will be able to customize their own
land use typology based on need and context. In this way, rather than building an
instructional approach of land use classification, the research seeks to develop a
compositional approach, a syntax of building the land use typology.
14
1.3
Framework of study
Generatedby the author
LITERITURE REVIEW
CONVENTIONAL LAND USE CLASSIFICATION SYSTEMS
RESEARCH QUESTION
HOW TO DEVELOP A VALID LAND USE TYPOLOGY?
THEORATICAL FRAMEWORK
THE
PACKAGES
THE
ELEMENTS
GENERIC
APPLICATIONS
c
0
THE PROCESS
>
r - - - - - - - - - - - - - ------- --
| Defined Package
- - - - - - - - - - -r-- - - - - - --
Defined Land
Use Elements
Defined Applications
CASE STUDY
Figure 1-1 the Framework of Study
The figure above (Figure 1-1) shows the framework of this study. This thesis will start
from the literature review on the development of land use practice. Based on that, I will
develop the syntax of composing a land use typology by identifying the key elements of
land use description. The syntax will describe land use typology as a specifically defined
package of these elements. It also includes the principles and process of building land use
typology.
While the syntax is an abstract concept, this thesis will propose several land use
typologies with purposes such as land use planning and transportation study. The
research will also propose generic applications of the land use typologies for land mixed
use, land use conflict, land use change and land use estimation.
With the theoretical framework, the professions and scholars might be free to build
specific land use typologies for their use. However, the feasibility of using these land use
typologies could also be limited by the limits of data. With this concern the research
15
introduces the new ideas and possibilities that come along with big data, expecting big
data will enable new land use typologies.
Then I will conduct a case study with Boston, U.S and Shenzhen, China, where a big data
based land use typology is built and tested. There are two purposes of the case study: one
is to illustrate the syntax of building a new land use typology; the other is to explore the
possibilities that big data meet the study of land use. In the case study, we will be able to
evaluate the syntax and the specific typology of land use.
Based on the framework, the first chapter of this thesis is the introduction of the
research. It identifies the research question, the framework of the study and the
terminologies that are used. The second chapter includes a literature review on land use
classifications, land use information system and land use, study, with a summary of the
involvement and problems. The third chapter describes the syntax of defining land use
typology, where the basic components, the process and applications of defining land use
typology. The fourth chapter describes the concept of big data and reflects on the
opportunities of big data. The fifth chapter describes the case study that is developed
with Boston, U.S and Shenzhen, China. The last chapter is a reflection on the value, the
limits and further researches of the study.
1.4
Use of Terminologies
I would like to clarify the key terms that are useful in this thesis and the relationships
among them (Figure 1-2).
Type/Class: a group of things that share similar characteristics and forms a smaller
division of a larger set.
Typology: according to Merriam-Webster's Dictionary, "typology" is "a system used for
putting things into groups according to how they are similar" and "the study of how
things can be divided into different types." By the term of typology, this research means
to study and classify land uses according to their features.
16
Classification: "the act or process of putting things into groups based on ways that they
are alike." In this research, I use "classification" as a result of dividing land use into
categories. I also use the term "land use classification" to refer the existing systems to
classify land use as a convention.
Syntax: the term of "syntax" is commonly used in linguistics. It means "the way in which
linguistic elements are put together to form constituents. A syntax is an orderly system
that arranges the components or parts harmoniously. In this research, I use "syntax"
specifically to refer to the principles, rules, components and other elements of
constructing a land use typology.
Generatedby the author
CLASSIFY
LAND USE
INCIDENCE
STRUCTURIZE
LAND USE
TYPE
COMPOSE
LAND USE
TYPOLOGY A
COMPOSE
THE SYNTAX
Land Piece A
Type
1
T
Elements
Land Piece B
Type 2
Type 2
Principles
Land Piece C Z
Type 3
...
Type 4
T
3
Type 6
T
4
TS 7
Context
Type 5
Figure 1-2 the Relationshipofthe Terms Used In This Thesis
17
LAND USE
TYPOLOGY B
2
2.1
I
URBAN LAND USE AND PRACTICE
U. S Land Use Classifications
The history of land use classifications and its practice has never been coherent. In 1876,
the U.S Congress laid down a suitability-based classification of public lands:
Lands arable without irrigation, lands suitable for farming with
irrigation, commercially valuable timber-bearing lands,
commercially valuable coal-fields, private lands, and lands
suitable for town sites.
In the afterward years, a different set of concerns rose beyond this system- for example,
the conservation of natural resources, conflicts between private construction and public
interests.
The former concern promoted the establishment of land use classes such as "natural
reservation." In the 1930s, when the American economy was in depression, the New Deal
programs began to classify agricultural land with a view of efficient utilization and soil
conservation. Farmland was classified as "marginal," "submarginal" or "super marginal,"
with an aim to reduce the first two classes and to reassign them to uses such as "forest" or
"recreation"(A. Guttenberg, 2002).
The latter concern contributed to the regulatory system that today we call zoning. When
first approved in New York in 1916, zoning as an ordinance divided land use into three
broad classes- residential, commercial, and unrestricted. These rough categories
established a foundation for a finer classification of land use. Later to implement a finer
policy of land use regulation, planners and officials expanded such rough categories into
a larger number of classes and subclasses, as exemplified in the categories used by the
Department of City Planning, the City of New York (Table 2-1). This system still serves
as a mainstream approach for zoning practice and employed by the cities' department of
planning.
18
Table 2-1 Urban Land Use Classificationfor Zoning
NYC Department of City Planning
Urban Land Use Classification
Residential Uses
One- and Two-Family Residences
Multi-Family Residences
Mixed Residential and Commercial
Commercial Uses
Industrial / Manufacturing
Transportation/Utility
Public Facilities and Institutions
Open Space and Recreation
Parking
Vacant Land
Source: NYC Departmentof City Planning
Table 2-2 Functional Uses of urban land, HarlandBartholomew
Total
Developed
Privately
Municipal
Area
Developed
Single Family
Two Family
Area
Multiple Dwellings
Commercial
Light Industry
Heavy Industry
Public &
Streets
Semi-Public
Railroad property
Use
Parks & Playground
Institutions-Cemeteries-Churches-City Property-
ETC
Vacant Property
Source: Urban Land Uses(Bartholomew, 1932)
As a combination of land use classification and management, zoning-based land use
classification system has become the "lingua Franca" of urban land use planning. In the
19
1920s along with the emergence of zoning, Harland Bartholomew, one of the first
planning consultants in the U.S, conducted a detailed survey of urban land uses that
included the entire city of Washington, DC. Thereafter, he launched a number of zoning
surveys around the country. In 1932, Harland Bartholomew finished his landmark study
Urban Land Uses, published by Harvard University Press in the City Planning series
edited by Theodora and Charles Hubbard. Bartholomew as a representative of urban
planners in his age, interpret land use classification as a tool for practice and classify
urban land use mainly as part of a scientific zoning survey (Figure2-1). In his land use
classification system (Table 2-2), he pointed out that the urban land naturally divides
itself into developed and vacant land. The former includes all of the areas that are used
for urban purposes. In this function-based classification, farming tracts built upon were
considered as in urban uses, and farming or truck gardening area often indicated as
vacant property.
Source: UrbanLand Uses(Bartholomew, 1932)
COMMERCIAL AREAS
W'OIUSRAA OAALMAD AREAS
URBAN LAND USES
JEFFERSON CITY MISSOURI
AILA XASIA
TWO
0FAMILY
AREAS
.'
1MAE*FAMILY AREAS
Figure2-1 Land Use Map by Bartholomew
Zoning based land use classification was widely recognized. However, it is worth
noticing that the nature of zoning is to designate permitted uses (Fgure2-2). Accordingly
the land use classification is designed to support the administrative purpose rather than
answering the question about "how the land is used?" or "What do we mean by land use
20
classification?" These questions were not posed in a systematic manner for several
decades of practice.
Source: NYC Departmentof City Planning
Figure 2-2 Zoning Map of New York City
In the 1950s, when economic of America was booming, the blueprint for the fast growing
metropolis is the most important issue for planners. The professions of planning started to
use computers with a greater capacity for data management and processing. Studies on
transportation and land use forecasting models were well experimented, such as the
Chicago Area Transportation Survey in 1955 and the Penn-Jersey Transportation Study
in 1960. Under this context, the rough categories such as "commercial" and "residential"
no longer sufficed. Planners and analyst complained about the chaotic state of land use
terminology(Sparks, 1958): its different names for the same use, same names for different
uses, and consequent inability to yield reliable quantitative data (A. Guttenberg, 2002).
The problems brought the classification of land use into the realm of scientific discourse.
Scholars, such as Robert B. Mitchell and Chester Rapkin, sought to understand the
"nature of metropolitan use and the forces that control it"(A. Guttenberg, 2002). Their
groundbreaking study explored the interaction between the urban land use pattern and the
21
underlying system of people/good movement. While in their 1954 book, Urban TrafficA FunctionofLand Use, they expressed the ambiguity of the term "land use":
Land use has many specific meanings. It may refer to buildings or
other improvements on the land, to the occupants or users of the land,
to the major purpose of the occupancy of the land, or to the kind of
activities on the land. (Mitchell & Rapkin, 1954)
For Mitchell and Rapkin, it was critical to answer the question "what is land use." They
chose to define land use as "the major activities of establishments based on the land."
Although "categories of land use should not be confused with categories of buildings,"
they decided to use the latter in place of land use in their traffic model(Mitchell &
Rapkin, 1954). Two years later, Mitchell's colleague John Rannells published his book
the Core of the City. Sharing a similar concept with Mitchell and Rapkin, Rannells went
beyond to take both urban activities and the physical environment into account.
The same year in 1956, the Planning commission of Philadelphia was invited by the city
to suggest suitable land use categories for a citywide real estate inventory. Albert Z.
Guttenberg was in charge. With an understanding of the previous studies by Mitchell,
Rapkin as well as Rannells, he suggested an inclusive approach to the problem of land
use classification, "land use was a concept with many meanings, many dimensions, each
dimension required its own distinct nomenclature"(A. Guttenberg, 2002). In this system,
he sorted out the multiple dimensions of land use, such as activity type, function,
development status (Table 2-3). The result was published in 1959 as "A Multiple Land
Use Classification System" (MLUCS).
In 1965, New Directions in Land Use Classification by Guttenberg was published,
wherein the idea is "to lay the groundwork of a grammar of land use planning"(A.
Guttenberg, 2002). According to Guttenberg, it was "found less acceptance" because of
"the costs of replacing or even, as in this case, just expanding an ongoing data system."
22
w
Developed, structure
Developed, structure
Developed, no structure
Undeveloped
Development Status
Multistory offices
Produce warehouse
Kiosk, pavement
None
Facility '3pe
Office activity
Storage and handling
Parking
Play
Activiy 23pe
Manufacturing
Agriculture
Retail trade
Recreation
EconomicFunction
Medium
Medium
Large
Small
Size
Glare
Sound, odor
Sound
Sound
Effect
More than local
More than local
Local
Local
Range
Activiy Characteristics
5
3
4
1
Time-Shape
Source: MultidimensionalLand Use Classificationand How it Evolved (Guttenberg,2002)
Source: Albert Z. Guttenberg, The LanguageqfPlanning:Essqys on the Originsand Ends qfAmerican PlanningThought (Urbana: University of Illinois Press, 1993).
Reproduced with permission.
Note: This table shows how a set ofbypothetical parcels can be multiclassified. Size refers to the number of persons using the parcel daily or for some longer period.Efect
refers to the type of impact the activity has on the surrounding area. Range refers to the spatial extent of the Impact. The numbers In the 'Tme-Shape column represent
hypothetical curves describing variation in numbers of persons and/or vehicles participating in the activity daily or for some longer period.
W
X
Y
Z
Parcel
Table 2-3 Multiple Dimensional Land Use Classification
With a similar idea of land use classification, in 1965 the American Planning Association
(APA) launched the Standard Land Use Coding Manual (SLUCM). Based on the
Standard Industrial Classification (SIC), the manual used a 4-digit hierarchical system to
identify and code land use activities. In 1993, APA Then, in May 1996, APA initiated the
Land-Based Classification Standards (LBCS, Figure 2-3) project to update the 1965
SLUCM. The first version of LBCS was released in 2000, with a purpose to allow
jurisdictions, agencies, and institutions at the local, regional, state, and national level to
share land-based data. LBCS has been widely disseminated through the APA Website
and is currently being tested by various governmental jurisdictions throughout the
country (Appendix. A).
Source: American PlanningAssociation
mom
MLOCS Acthvity
0
LBCS Function
UmEJ
M
]LBCS
Structure
L*CSSite
LBCS Ownership
Figure 2-3 APA Land Based ClassificationStandards
2.2
China Land Use Classifications
China did not carry out a systematic land use classification until the 1980s, when the
State Bureau of Land Administration was established, and The Law of Land
Administration was launched. Since then, China has been following national wide
standards of land use classification.
In 1984, Chinese Committee of Agricultural Regional Planning laid down the Technical
Rules of Land Use Survey. The standard classified land use into eight primary
categorizations: cultivated land, garden land, forest land, grassland, land for residential
areas and mining, land for traffic, water body and unused land. The primary categories
24
were divided into 46 secondary categorizations, according to what local government
could develop more detailed land use classification system as needed. In 1989, the State
Bureau of Land Administration distributed the protocol for urban land survey (TD10011989), classifying urban land into 10 primary categories and 24 secondary categories.
With a purpose of integrating urban and rural land in land survey, the two standards
mentioned above were combined as a trail National Land Classification ([2001]255),
classifying the land into agricultural land, land for construction and unused. This trial
standard was then developed and launched as the national standard of land classification
in 2007 (GB/T 21010-2007). This standard has 12 primary categories (Table 2-4) and 57
secondary categories (Appendix. B).
Table 2-4 China NationalStandardof Land Classification
Primary category
Main type
Code
Name
1
Cultivated land
2
Garden land
3
Forest land
4
Grassland
5
Commercial
6
Mining and warehouse
7
Residential
8
Public services
9
Special use
10
Land for transport
11
Water body and facilities
Development land / Agricultural land / Unused
12
Others
Development land / Agricultural land / Unused
Agricultural land
Agricultural land /Unused
Development land
Development land / Agricultural land
Source: the Ministry of Land and Resources of China
In urban areas and the field of planning practices, China has its Code for Classification of
Urban Land Use and Planning Standards of Development Land (GBJ137-90) since 1991.
Through a classification system and the planning guidelines for each class of urban land,
this Standard works as a technical reference for master planning and urban development
25
of Chinese cities. In the implementation process of the standard, each city can adjust its
land use standard according to its development conditions. As China is undergoing rapid
urbanization process, the development condition around the cities demonstrated all kinds
of diversity. A variety of new land use patterns also begin to appear in some cities, which
makes the "standard" gradually in some areas become out of date (Wang, Zhao, & Li,
n.d.). For this reason, in 2010, the Ministry of Housing and Urban-rural Development
enabled the new standard (GB 50137-2011), with multiple factors involved determining
the classes of land use. The new standard classified urban and rural land into two primary
categories, nine secondary categories and 14 sub-categories (Appendix. C). Among the
14 sub-categories, urban construction land is subdivided further into eight primary
categories, 35 secondary categories and 42 sub-categories. This code has been serving for
planning professions as a national standard.
2.3
Land Use Information Systems
In most cases, urban land is under planning regulation or institutional management. For
these purposes, land related information need to be registered and structured at a certain
spatial level. Planners often breaks it into ownership units or parcels because these are the
market units upon which development projects and land use changes happen (Kaiser,
Godschalk, & Chapin, 1995).
In the U.S., when creating land use information module, land is usually seen as functional
space devoted to various uses. Activity information are considered, because they are the
patterned ways where households, firms and institutions act in their daily affairs in urban
areas. However, for practical reasons, the land use information system is often designed
to focus on more stable characteristics. The reason is practical- while traditional land
use survey often takes a long time to perform, land use activities data is difficult to obtain
and might only be valid for a short period.
As a result, a framework of land use database often includes inventories of land supply,
land policy and activity systems (Kaiser et al., 1995):
26
Land supply inventory-includes the existing and projected supply of
developed and developable land. It contains information about the
nature and physical condition of the built environment, as well as
land availability for development of different urban land use type;
Landpolicy inventory-includes existing regulations, procedures,
plans and policies that affect land use in the local context. It often
obtains relating information from a variety of government agencies at
the local, regional, state and federal levels. Zoning regulations are
typically included in this part.
Activity inventory-includes urban land use activity patterns. It
includes aggregated records of the journey-to-work commuting. This
data has been gathered through surveys, interviews, observations and
mechanical counts (Chapin and Kaiser 1979, ch. 7) and mainly
supported transportation research. Activity information supports
planning decisions about the physical settings of urban areas.
In China, the land use information system has been under rapid development. In 2011,
the Ministry of Land and Resources adopted its initiative in developing a national-wide
planning information system. The system will be based on the second national land
survey database, follow the standard for land use planning, and service as administrative
tools.
Automated identification and characterization
After the 1960s, remote sensing technologies was widely used and after the 80s,
significant progress was made in the development of new remote sensors with fine spatial
and spectral resolutions (Hu & Wang, 2013). Remote sensors are able to capture the
physical characteristics of land use. Accordingly, researchers developed the techniques of
automated urban land-use classification. The way it works is that researchers first
identifies individual land-based attributes as they become relevant and then assign the
attributes to its fields. With the help of computer manipulation, we then can combine data
from remote sensors with the attributes, and derive information for classifying land use.
27
The remote sensing techniques can assign classes of land use to pixels based on their
spectra (Fisher 1997; Lu and Weng 2006), textual (Myint 2001; Shaban and Dikshit
2001), or contextual properties (Gong and Howarth 1990). Some researchers also
conducted studies based on the spatial unit of fields instead of pixels.
The use of remote sensing and automated land use classification techniques has largely
improved the efficiency of collecting land-use information, as well as maintaining the
land use database updated. Nevertheless, there are still some land use classes calling for
direct separations from each other, such as office, civic, industrial, and transportation
land (Wu et al. 2009). More critically, this land use classification mostly relies on the
physical condition of the built environment, therefore, rarely able to derive finer
differences of land use classes. Now, as mixed-use is more and more emphasized,
specifying the differences between two parcels will be more difficult. After all, the
automated land use classification system is based on a land-cover difference of land use,
rather than the activities and uses of urban land.
2.4
Land Use Analysis
Based on the land use information system and availability of data. Two types of land use
analysis are well developed.
Developabilityanalysis: the Developability of land is its capacity to
be put to urban uses. In the simplest meaning, developable land is
vacant land without server physical constraints that could be planned
or zone for more intense use. Developability analysis screens the
land supply in order to locate areas suitable for future development
or redevelopment (Kaiser et al., 1995). Three techniques can be used
for developability analysis: suitability analysis, carrying capacity
analysis and committed land analysis.
Imageability analysis: imageability analysis was initiated by Kevin
Lynch. Planners use this method to understand how the residents and
28
visitors view the city. Image information could be collected through
interviews and surveys.
Since the 1950s, land use has been considered as a crucial factor of land-transportation
interaction in the U.S. The mutual relationship between transportation network and land
use patterns is a well-accepted concept (Iacono & Levinson, 2011). The physical location
of the transportation network can exert a strong influence on patterns of urban settlement
and activities (e.g. Figure 2-4). In turn, the features and spatial distributions of activities
will inevitably influence travel behavior and transit system. However the conceptual
model of interacted land use and transportation has been suffering from the confused
definition of urban land use and its vague focus on activities. As a result, the term of
"land use" in the transportation field and practical planning field often means different,
the former focus more on activities and behavior, while the latter could provide
information more about ownership, economic functions and development status.
While in China, the scholars have been focusing two aspects. One is the trends and
patterns of land use change, which relates to both urban and natural geography. The other
is about land use assessment and evaluation, with respect to land use efficiency and
intensity.
29
Source: Livable Streets (Appleyard, Gerson, Lintell, & more, 1982)
'intre~7
*
3T
7r
LIGHT TRAFFIC
2000 w..e.,
,
200 VVad1s 01
d,"
Pass
) fomnda aw
3 Csiine
en~~~
h
-
C
.
MOOETE TRAFFIC
O000 Vebmcas pat day
w0 *B..
w
-ook
Pot Patton
1146ft" -
ft"
J
iens
0.9 qrndes
*WAVY TRAFFIC
Is000 Ve'6$ O pOft
VOW0
V~alml
pea-
ero
Ow
*
3.1**~t6e6
ar
l
)
it's
cffr.t
t1
a
frr
.-.
not
t
t
.
I
~~rcpl
tc ,
v-v
'
I~t
am,
9Z
ar
t
,t
t~
tt.&
r~
Figure 2-4 the Interactionof Land Use and Transportation:Impact of Traffic on Activities.
30
traffic
afI~
rad~c
2.5
Land Use Change and Estimation
There are three general methods developed for forecasting land use change. Among
them, the simplest types are stochastic models (Brown, Pijanowski, & Duh, 2000;
Levinson & Chen, 2005), which treats land use change as a stochastic process and the
rates of change constantly before reaching equilibrium.
The second types are regression models. With endogenous and exogenous variables,
regression models can demonstrate physical and social influences on urban development
(Conway, 2005; Verburg, Eck, Nijs, Dijst, & Schot, 2004).
The third types are cellular and agent-based models. This method takes advantage of
advances in computational power and data storage. The models desegregate urban space
as individuals or land parcels, and simulate urban development based on the cellular
automata framework (Jantz, Goetz, & Shelley, 2004). Cellular automata models
emphasize more on dynamic interactions between agents and have recently gained
greater acceptance as tools for simulating land use change in urban areas (Iacono &
Levinson, 2011).
2.6
Brief Summary
In the history, land use classifications have been more of the product of expediency rather
than of rigorous thought (A. Guttenberg, 2002). And, the practices of land use
classification have been implemented by various agencies and branches of government
with different purposes, which brings inconsistency and difficulties to integrate the
information.
The involvement of land use classification systems has been in accordance with the
practical needs and technologies to enable the possibilities. There has always been a
mismatch between the growing need for detailed land use data and the limitation of
available tools. Although the framework of multiple dimension land use classification
was proposed for decades, the activity dimension of land use was not yet well interpreted.
So to implement the theoretical framework into practice is still very difficult.
31
Albert Guttenberg, the pioneer of multiple dimension land use classification, points out
some of the reasons in his recent article, MultidimensionalLand Use Classificationand
How it Evolved: (A. Guttenberg, 2002).
a. The discovery of a mismatch between a task in hand and the instruments
necessary to perform that task (Haig'sperception that traditionalland use
terminology was too crude to serve a study of the land use changes that were
occurringin the New York metropolitanregion).
b. The problem itself must be something of a "hot issue ", ensuringthe
continuing interest of a community sufficiently large and influentialthat sees
itself benefitingfrom a solution.
c. A fortuitousfactor,perhaps,but an important one is the ability of one kind of
problem to "piggyback"on another (Mitchell and Rapkin and Rannells were
not seeking to construct an improved land use classificationsystem, but their
work didfurther that end). In fact, leads and clues may be found in the
apparently unrelatedwork of even distantpredecessors (Hurd,for example),
which indicates why a knowledge of history can be importantfor scientific
progress.
d The existence of a resource base ready to provide the wherewithal to support
researchrelevant to the problem, meaning not only financialsupport, but the
bureaucraticcover anda protectednichefor the exploration of approaches
not in conformity with conventionalthinking. This is what the Philadelphia
City PlanningCommission provided in this case, thanks to the encouragement
of innovative staffwork by Arthur Row Jr.and Executive DirectorEdmund
Bacon.
e. Bringing to bearparadigmsfrom anotherfield (asfrom thefield of language
studies in this case) may help clarify a problem and suggest solutions not
otherwise available.
32
f
The cogency of a proposed innovative solution is necessary, but may not be
sufficient to win attention and respect. It helps immensely ifthe innovation is
seen to emanatefrom a source regardedas authoritativeand of high prestige
(the PhiladelphiaCity PlanningCommission,for example, occupies a
leadershipposition among city planningagencies).
In summary, the researchers of land use classification have been focusing on the practical
side and what should the land use classification look exactly like. However, there is no
researcher revealing the "black box" of building land use classification system. That said,
it is not clear on what should be a systematic method of developing a land use typology.
With this concern, this thesis seeks a common language of building land use typologies.
That is the syntax of land use typology.
33
I
3
3.1
THE SYNTAX OF LAND USE TYPOLOGY
Understanding Land Use
Despite with some variations, all of the classes of land use in different countries include a
wide spectrum of lands, such as forestland, water body, cropland, and developed land.
Yet with an intent to differentiate land cover and land use, the research here argues that
the land could be covered with forest, grass, wildlife, but could only be "used" by
human activities. It is human activities that drive social changes, economic growth and
urban change.
Human activities contain both virtual and spatial aspects. For example, working involves
both spatial movement (such as commuting) and social relations (employment and social
production), which reflect respectably the virtual aspect and spatial aspect of land use.
For land use study, I will not include the virtual aspect of human activities. Rather, I
consider land use as the spatial result of human activities (Figure 3-1).
Generatedby the author
Virtual
Human Activities
Spatial
Figure3-1 Land Use as a PhysicalDimension ofHuman Activities
34
3.2
The Elements of Land Use Typology
This thesis defines five elements as the key descriptive factors of land use typology. They
are the basic components to build a land use typology. Figure 3-2 is an illustration of the
elements applied to the land.
Generated by the author
Intensity
Function I
m
.nect
Fuinctio
3
Figure 3-2 the Elements of Land Use
3.1.1
Function
Function of land use generally means what people do on the land. This element of land
use has been widely used in the existing land use classification systems. For example,
the land use classification used by NYC Department of City Planning (Table 1) is could
be considered as a classification system that is mainly based on land use functions.
Dividing land use functions into classes such as residential or commercial is one of the
most common classifications of land use.
35
When building a land use typology, the definition of the function could vary case by case.
Function is an element along any scale of space of time. While the city could have its
function as the capital city, manufacturing center or transit hub, a parcel of land could
work as a recreational center of a community, landmark or construction area.
The concept of function could also be defined in the aspects of economics, social or
physical development. As an example, the economic functions of production could be
divided as primary (e.g. agriculture, mining), secondary (manufacturing) and tertiary
(retail, IT, service).
The function of land use is relative, because it relates to the actions of land use
contributing the larger scope. When we identify the function of a parcel is residential, it
implies that this piece of land generally provides more living accommodations than
parcels that are identified as commercial.
When developing a land use typology, the definition of land use function could be very
detailed or rough, according to the need. A very detailed division of function does
necessarily contribute to a better land use typology. For example, land use classification
for national level land resources management needs not a specification between single
family residence and multi-family residence. Likewise, a land use typology that does not
separate offices and residence is not good for zoning.
3.1.2
Intensity
Intensity means the strength of use. The intensity of land use could be estimated using
various indicators. For example, FAR indicates the physical intensity of the built
environment, whereas GDP could measure the economic intensity of land use.
The element of intensity is sometimes used in existing land use classifications. For
example, the classes of developed area and undeveloped area is a simple division based
on the level of development. In more cases, the information of land use intensity is
implicated or combined with the function of land use. For example, Table 1, under the
36
category of residential uses, the division of single family residence and multi-family
residence indicates that the former type is of less intensity of land use than the latter type.
3.1.3
Connectivity
Connectivity reflects how the land pieces being connected with each other. As early as in
1929, a writer Frigyes Karinthy published a story called "chains" in his book Everything
is Different. By this literal piece, Karinthy raised an argument that the world is getting
socially smaller. He claimed that people are increasingly connected and that the
interconnected world make everyone on Earth at most five acquaintances away from
anyone else. 85 years later, this is more and more the case by both physical and social
means. The world is more and more connected. After that when scientific interest was
raised in the study of networks, connectivity became an important measure of the
robustness of the network.
If we apply this concept into our land use study, we may find interpret it by its quality,
degree or capability of the land connection. We could define the connectivity of land
use according to the trade or migration, commuting, information exchange or other social
movements. We could also use identify the centrality of place with the measure of
connectivity-
a well-connected area in the city is possible to become the urban center or
sub-centers.
The concept of land use connectivity is similar to accessibility in transportation study.
When including connectivity in the development of land use typology, we will be able to
relate the classes of land to transportation study. As another example, walkability is a
measure of how friendly an area is to walking. It is also a measure of the quality of land
use connection.
3.1.4
Probability
Probability is the likeliness that an event happens. I consider the use of land as a fact of
probability in two senses:
37
First, probability of land use means the frequency or say, the proportion of certain land
use type. The proportion of land use could be measured by time or space. As an example,
the total area of a land parcel could be divided into 30% as retail space and 70% as
residential space, or a piece of land is half day for working and half day for recreation. In
more cases, the probability of land use refers to both spatial and timely variation.
Secondly, probability of land use could be understood as the confidence of some certain
thing is to happen. This interpretation of land use probability relates much to the
estimation of land use, which will be discussed later.
The two senses of land use probability, although with different focuses on the current
condition and the future, are much related to each other. While the frequency aspect of
probability represents the observation on the history or current condition, the confidence
aspect of probability is based on the baseline of the observation.
Generatedby the author
Figure 3-3 the ProbabilityField of Land Use
Based on the observation, the probabilityof land use feature is higher in some
places and lower in other places.It provides a baselinefor land use estimation
The probability of land use could indicate the function, intensity, connectivity or other
features of land use. With this concept of probability, we could imagine the defined land
38
use feature has a "probability field" in the city, which is generated from the observed
occurrence and serves for future prediction. The concept of "field" here borrows from
physics - it is a quantity that has a value for each point in space and time. Further, both
mathematics and physics have theories and application about probability that we could
borrow to implement in urban studies.
Source: Gonzdlez, Hidalgo, & Barabdsi, 2008
a
C
W
100
0 DI
700
02
10
600
Or + &rb)Pe-"
-
10-2
500
* 400
10
300
4
200
10-5
100
10-7
100
100 200 300 400 500 600 700
Di1tance (km)
-
-
101
102
104
10
Ar (km)
El10-3
4
10-2
E10-
-1 bIJ
-15
-15
b
0
x (km)
15
15
-150
0
x (km)
150
-1,200
0
1,200
10-4
x (km)
15
15
10-5
10-6
-15
-15
U
X/1
10
-10
-15M
15
-15
-
U
X/t,
0
I
15
X/01
Figure3-4 an Example Study on the ProbabilityofHuman Trajectories
Basic human mobilitypattern (upper) andthe probabilitydensity of
human behaviorpattern (lower)
39
mIn
3.1.5
Scale of Time and Space
Historically, geography is true to its name; it studies the world, seeking to describe, and
to interpret, the differences among its different parts (spatially), as seen at any one time,
commonly the present time (Hartshorne, 1939). However, in modem sciences, space and
time cannot be separated from each other (Figure 3-5). For example, in modem physics
theory, the uniform concept of space-time combines space and time into a single
interwoven continuum and time cannot be separated from the three dimensions of space;
whereas, in computer sciences, the merits of the algorithm are usually defined by both the
space complexity (the memory space that it takes) and time complexity (necessary time
for running the algorithm). That means at the same level of complexity, shorter time of
processing a task means more memory space to take, and vice versa. An algorithm could
complete more complicated tasks by either improve the space or time efficiency, or both
of them.
Generatedby the author
Time
Space
Figure3-4 Activities with the Continuity of Time andSpace
It is worth thinking that in both physics and computer sciences, time and space are highly
influenced by each other. This rule also applies to our social life at a certain level. Think
about the mobility of people, are we not spending time to move through space? Although
40
the progress of science and technologies has greatly facilitated our speed of movement,
we are still limited by the tradeoff of time and space.
Generated by the author
Time
Space
Figure 3-5 the Movement ofIndividual in Time andSpace
Hagerstand and his ideology of time-geography opened the first of its kind on the
geographical interpretation of time and space (Figure 3-6). At one level of analysis, timegeography deals with the time-space "choreography" of an individual's existence at daily,
yearly, or lifetime (biographical) scales of observation. Time and space are seen as
inseparable. Each and every one of the actions and events which in sequence compose the
individual's existence has both temporal and spatial attributes (Pred, 1977). Hagerstrand's
call for a time-geographic focus on people and, in particular, the event sequences which
constitute the days and life of each individual person stems from a humanistic concern
with the "quality-of-life" and everyday freedom of action implications for individuals of
both existing and alternative technologies, institutions, organizations, and urban forms.
However, time-geography cares more about individual activities, rather than the aspect of
land as I focus in this research.
41
Although time and space are continuous, we need to interpret the elements of land use in
a certain extend. I use "scale" to describe with large approximation the extent of the
time period and the size of the area studied.
Generated by the author, the maps werefrom New York Times
and other online sources
24 hours
30 days
12 months
Longer
Figure3-6 Scales of Space and Time
As a variable, scale indicates the level of detail and specificity of land use information,
which increases as the level of concern descends spatially (e.g. Global, county, region, to
city, neighborhood, and parcel level) or timely (e.g. Decades, year, month, day, hour,
minutes).
In the past, when map could only be drafted by printing, a certain map might be presented
at only one spatial scale - 1:50000 means one centimeter on the map represents 500
meters in reality. "Large scale" refers to maps on which objects are relatively larger than
maps at "small scales". However, with the limit size of a map, larger spatial scale also
42
means a smaller area represented. Under that condition, spatial scale is a tradeoff between
the size of the concerned area and the specificity of information.
And because of the difficulties in the land use survey, the land use maps often have a
very approximate time period. For example, in a map called Boston 1995 land use map, it
is difficult to know when exactly the data were collected, and more importantly, in what
timescale it is the case.
With computer-assisted mapping programs and new data sources, we can collect data at
any desired scale consistently. Ideally information collected is detailed enough to be
transferred to any scale. Then for scale, we will only need to consider what is the
appropriate level of specificity or aggregation of data. The concern is based on the simple
facts that we like to know the variations about land use, and that our cognitive system can
hardly perceive information outside the scale of our focus-
looking at a neighborhood
map, we want to know the exact location of a primary school, but looking at a global
map, a single school matters little if not nothing to us.
3.3
Land Use Typology: a Package of the Elements
Generatedby the author
POSSIBEL INTEPRETATIONS
FUNCTION
INTENSITY
THE ELEMENTS
0
4
PROBABILITY
CONNECCTIVITY
SCALE
Figure 3-7 Land Use Typology as a Package of the Elements
43
"
As function, intensity, probability, connectivity and scale are considered as the key
element of land use, each of these elements allows various definitions according to the
context of use. A defined land use typology, in the syntax, will be a package where each
element is defined concretely and in the proper order.
The syntax allows the many possibilities to define and combine the five elements. With
the scale as a mandatory element to be included, certain land use typology does not have
to cover all of the five elements. It could happen that a land use typology includes only
some of them, or a single element is defined by two means.
Based on the purposes and uses, the defined typologies could vary from each other. This
thesis will discuss the valid packages for different aspect of general use of land use
typology.
3.3.1
Packages for Energy Analysis
By a broad definition, energy consumption is a result of human activities and land use.
The study of land use typology in the context of energy use will improve our
understanding on energy use, economy and climate change, as well as elevating out
ability to mitigate climate change.
Table 3-1 an Interpretationof the Development for Energy Based Land Use Typology
Elements
Function
Implications and measurements
Mandatory
E.g. Income and household conditions, building type, energy facilities
and appliances
Intensity
Mandatory
E.g. Energy consumption
Probability
Optional
E.g. Energy structure, proportion of the functions
Connectivity
Optional
E.g. Supply and demand relationships
Scale
Mandatory
E.g. Nation, city, district, block, building
Generatedby the author
An energy based system of land use typology focuses on the aspects of land use that
could reflect or influence the status of energy use. As an example (Table 3-1), function
and intensity could be the basic elements that are involved, with the function as an
44
indicator of the energy use forms and intensity the strength of energy utilization.
Probability and connectivity could be optional, which will be useful in energy estimation
and supply-demand analysis.
As energy consumption could be analyzed in national, city, neighborhood or building
level. We should develop the elements accordingly so that the measures are the best fit of
aggregation and specificity.
3.3.2
Packages for City Planning
Planning is a process to ensure the orderly development of settlements and communities.
Planners and decision-makers have been the main users of existing land use classification
systems. With the syntax, planners could develop the land use typologies based on the
existing measures and standards of city planning.
While scale is an essential element of planning, the purpose and focus of city planning in
different scales vary from each other. The corresponding land use typology should be
defined according to the context, such as the purpose, the unit of planning, the
institutional context or the related planning scheme. As an example, the table 3-3 shows a
possible interpretation of land use elements according to the scale of planning.
Table 3-2 an Interpretationofthe Land Use Typology Elements in Different Scale of City Planning
Scale\
Function
Intensity
Probability
Connectivity
Regional
Designated function
GDP of the
Industrial structure
Regional corridors,
scale
of city
city
economic connections
Overall
Urban transportation
City scale
Primary categories
Examples
District
disrbution
distribution
network
Proportion
scale
of function
P
and intensity
Secondary categories
Street network
FAR
Parcel
Detail categories,
scale
existing regulations
Building connection
Generatedby the author
45
3.3.3
Packages for Land Resources Management
We should notice that with this syntax of land use typology, the classes that are
developed could be very simple. For example, since the land use typology for land
resource management needs to cover the whole area of a country or region, a less detailed
categories of land use will serve better for the strategic layout of land resources (Table 33).
Table 3-3 an Interpretationof the Developmentfor Land Resource Based Land Use Typology
Elements
Implications and measurements
Function
Mandatory
E.g. By suitability, development/non-development, rural/urban
Intensity
Optional
E.g. Degree of development
Scale
Mandatory
E.g. National, regional, city
Generatedby the author
3.3.4
Packages for Transportation Study
The coordination of land use and transportation has been recognized widely since the
1950s, where transportation is considered as a result of the activities on the land and the
process of influencing land use change. In the past, transportation researchers have
developed the necessary measures of land use based on their need, but the transportation
based land use typology was not fully studied or applied in land use practice. Table 3-4 is
an example to interpret the elements of land use typology for transportation study.
Table 3-4 an Interpretationofthe Developmentfor TransportationBased Land Use Typology
Elements
Implications and measurements
Function
Optional
E.g. Income and household conditions, behavior pattern
Intensity
Optional
E.g. Frequency of transit
Probability
Optional
E.g. Share of mode
Connectivity
Mandatory
E.g. Public transit network, road network, street network
Scale
Mandatory
E.g. National, city, district, neighborhood
Generatedby the author
46
3.4
Generic Applications
With various purposes and focus, the developed land use typologies are applicable and
adaptable for many purposes. The applications, which are applicable for any developed
land use typology, emphasize different elements of land use features (Figure 3-9). While
this paper proposed three generic applications, they are not the only ones that are
possible.
Generatedby the author
F353
NEN]
PROBABILITY
CONNECTT
E
LAND USE
CONFLICT
CHANGE
N
ELEMENTS
ESTIMATION
/
OTHER
APPUCATIONS!
SCA LE
Figure 3-8 the Various Focus ofthe GenericApplications
3.4.1
Land Mixed Use
Mix use focuses on the function and probability of land use. It means different
functions of land use happen in close proximity to one another. Mixed land use is a
critical component of smart growth in achieving better places to live. It is considered to
provide a diverse population and commercial base for supporting viable public transit,
and enhance the vitality and security of the area.
47
The concept of mix use defined in this research has two dimensions. One is the spatial
dimension of mix use: the functions could be physically close to each other. An example
is the multiple functions of a single building or parcel. Another dimension of mix use is
the temporary aspect of proximity. For example, a place has a different use at daytime
and at night. In most cases, the land is mixed used in both of the two aspects.
In modem time, cities grow larger to facilitate the subdivision of labor and urban land
often demonstrate a mix of use. A place could be characterized by a specific function, for
example, a park, transit station, a governmental center, but the place could conduct other
functions and services such as retailing in a park. We could be confident that land use is
more or less mixed in functions because of the nature of human activity. However, this
does not prevent us from identifying land use with a
3.4.2
Land Use Conflict
A land-use conflict occurs when there are conflicting results on the land, such as when an
increasing need creates competing demands for the use of the land, causing a negative
impact on other land uses nearby. I consider land use conflict as a combined focus on
land use function and land use connectivity.
For example, Figure 3-10 is a combination-graphic of a correlation table of land use. The
conflict types of land use are indicated by alternating shades of gray. The correlation
table relates the 49 conflict issues (along the left) to the 164 conflicts, arranged according
to conflict type (along the top). Whenever an issue is present within a conflict, the
corresponding cell in the table gets marked in black.
In urban areas, two common types of land-use conflicts in urban areas are residential,
industrial or residential-transport land-use conflicts. This is due to the noise, air and water
pollution created by the inappropriate function that are not separated from the residents.
With the information about land use functions and their connections, the planners will be
able to identify the land use conflicts and separate land uses that do not complement each
other.
48
Source: von der Dunk, Gr&t-Regamey, Dalang,& Hersperger,2011
COMS
me-""
n~a..am. aawax
"
%-a
...m.e
-
am
~in1dp~i.hIn.
I
lad
ransm" efrlfd
imsbiatmsu=w
.o-re in.- . aId.
Vi."
sil
U-smn
--
0
.uk
Visualp.re.a
U
iMOW*MA b O SM )
~madrv..l~ensaks.
-.1 (doww w ates a
,.am...urin.m.dalrn-sa-am)
la ,
a.s(-"--.1#8.
-ei--n-s.
Issaf---oma
- e..
rnsbrb
l..mmi---m.am..)
Uai..mo.4modo
bdbm~keb
Is
0.
ILB=4p~~jLr~'i~
Figure3-9 the Identificationof Land Use Conflicts.
49
3.4.3
Land Use Change and Future
Land use change is a process that the use of land is transformed over time. It relates to
every aspect of land use. In the urban area, the driving force of land use change is often
human activities.
With land use changes continuously, land use estimation is a prediction of land use
future. It relates to the probability of land use function, land use intensity and land use
connectivity. It is based on the implication that the elements of land use a probability of
happening at a place and from that we could predict the likeness of the functions
happening in the future (Figure 3-11).
Generated by the author
Figure3-10 the Prototype ofLand Use With ProbabilityField
Machine learning, a branch of big data and artificial intelligence might be helpful for land
use estimation. With machine learning, the system is trained from input data. After
learning, the system will react accordingly for the future conditions. Currently there are a
wide variety of machine learning applications, such as game playing and information
retrieval. But it has not yet been used in the field of urban studies.
50
3.5
The Process of Application
3.5.1
The Principles
The syntax establishes a framework to develop a new land use typology. While
practically developing a new land use typology, some principles should be considered
during the process.
Reliability and validity: the new typology should be reliable and valid. In other words,
accordingly the criteria for a theory model would be: (a) at what level of complexity or
simplicity it would be, in other words - the balance between detailed describing the real
world and efficiently presenting what we need to know; (b) in what extent it is true. In an
ideal condition, the information kept should be the most efficient and relevant.
The significance and necessity: the syntax of land use typology allows numerous
packages of defined elements. However a land use typology is only valuable if it solves
problems that the current system could not solve, the solved problem is significant
enough to develop a new land use typology.
Measurability and operability: the new system should be operable and feasible to adapt
in practical condition. Measurability relates to the accuracy and clarity of the defined
elements. The availability of data is often the constraint of developing an operable land
use typology.
Adaptability and flexibility: the new typology should provide a space for adaption
under imperfect condition, such as insufficient data and no official data. For example, the
new typology of land use, although as a different framework for land use description,
could make a one-to-one correspondence with the classes of conventional land use
classification. The purpose of this principle is to make the new framework adapted to the
established ones and to enable the existing data transformed into the new system.
51
3.5.2
The Unit of Measure
When establishing a measurement, a key question is what the spatial unit of measurement
is. In the past, the problem was focused on the tradeoff between accuracy of small unit
and the difficulty to get data. For example, whereas with the capability of getting data
largely improved, the unit of measurement in current times could be in the "pixel" level.
We should think more about what is the best unit to make meaningful sense of the data.
We assume that through big data, the main data sets to deal with are geolocated points
indicating an individual activity or person. This is for two reasons: 1) for practical
reasons, the format of location-based big data are often venues or somebody at a place. 2)
Considering human activities in the city, the "use" of the land is based on individuals,
which are often seen as "points" in the city. I will not identify those "points" like a piece
of land. Neither could any act about land happen in this scale. Therefore, I will need to
aggregate the possible data source before implement the measurement.
Accordingly, three unit systems were considered: grid based unit, data based unit and
object based unit.
a.
Grid based unit
Source: QGIS Introduction
Source: Generatedby Liqun Chen and QianqianZhang,MIT
Raster
Rate
Pixe
U
FFM
__
columns
Figure 3-11 Raster Unit anda Raster BasedHeat Map
52
A grid unit system is a framework of areas that are divided by intersecting lines. In the
U.S., the grid unit system has been used for land surveying under the Public Land Survey
System (PLSS), where a section is an area of nominally one square mile with 36 sections.
The advantage of the grid based unit system is that the unit is universally consistent. And
the grid could be as large as a city, and as small as a pixel. Therefore, the grid based unit
system could deal with many types of analysis. The disadvantages are that the grid is not
reflected on the external features of the land, such as land use boundary.
I consider raster cells as a grid system, because the defined units of measurement
distribute evenly in the study area, and the cells are in the same size. The only difference
between an ordinary grid system and the raster cells is that each raster cell could include
data for its surrounding area. The reason is simple: as the raster cells are very small,
expanding the sampling area will make the result a smooth surface. Otherwise, the
variance of point data will make the result difficult to read.
b.
Data based unit
Source: ESRI GIS Introduction
-I-...........
Input point coverage
Thiessen polygon
coverage
Source: Panda Whole website
.. ..
TIN
Bisedted TIN
Figure 3-12 Thiessen Polygon And an Example Map
Instead of dividing spaces evenly, we can also divide the space into a number of regions
based on the input data. By this method, each unit defines an area of influence around the
sample input data. Voronoi diagram (also referred as Thiessen polygon) is a typical
53
example (Figure 3-13), where for each seed (input sample point) there will be a
corresponding region consisting of all points closer to that seed than to any other.
The data based unit system is suitable for analyzing the radius of influence. But since the
result is highly dependent on the input data itself. It is difficult to build a standard land
use measuring system based on this.
c.
Object based unit
The object based unit system means that the geographical area is divided systematically
for administrative or social reasons. There are many practical examples of this unit
system. For example, building footprint is an object based unit system with a focus on
buildings (Figure 3-14); the census geographic entities are designed for systematically
acquiring and recording population data; land parcels are lots with defined boundaries
and often serves as the units of land administration and management.
'ource: Boston Redevelopment Authority
-W
I
IV
-
-15
,
CityHail Plaza
............
Figure 3-13 Building Footprint as an Object Based Unit System
54
The object based unit system is naturally related to the feature of land (such as
ownership, street network), therefore it is in wide use of land use interpretation.
However the division of land areas of this system has been often a combination of
historical reasons and practical convenience. And the boundaries of land units may
change over time.
3.5.3
The Steps
Generatedby the author
Interest
Define the Elements
Activity
Transportation
Energy
Built
Environment
Scale
Function
Intensity
Probability
Connectivity
Pic k the Unit
Grid based
Object based
Data based
of Measure
Data mining
Application of
First hand data
the Typology
Official data I
Current
Open source data
...-.
Condition
Land Use
Change
Land Use
Predition and
Intervention
Figure 3-14 the Steps of Land Use Typology Development
To apply the framework of new land use typology, several steps need to be considered
(Figure 3-15). While the steps are organized linearly, the steps are influenced by each
other in the practical condition.
Identify the topic of interest/Purpose: as discussed above, the syntax of land use
typology could be applied to different topics, based on what the elements and packages
could be accordingly defined. The topics related to land use includes, but not limits by
energy consumption, transportation and activities, city and regional planning, real estate,
etc.
55
Choose a scale and the unit: scale and units are the fundamental variances of land use.
The variance in scales and spatial unit will result to different definitions for each element
of land use typology. As space and time are continuous, we could pick any scale and unit
of measure. However unseasonal scale and unit, such as 4/3 days or 0.5 buildings, will
lead to a difficult process and unintelligible results.
Define the elements: Based on scale and the topic of interest, we could define the
elements of the land use typology and its applications with specific measures.
Relevant datasets/data mining: data are essential to apply the theory into practice. The
availability of data needs to be considered at the beginning of the study. And every step
of applying the syntax is affected by the quality of data. With the development of big
data, three types of data sources might be useful for land use study: (1) venue data or
information about places. This is the type of data that is used in the case studies. The data
could come from social media, yellow pages, government data or agency surveys; (2)
Individual or transaction data are information about a single act or a single person's
trajectory. The data could come from mobile phone traces, taxi record or car trajectories;
(3) Open source maps and other crowdsourcing platforms bring together different types
of data. This includes services such as google map or open street maps who integrate
information from the users.
Interpretation and presentation: the traditional media of 2D map has limited the
volume of information to be delivered. Now with new tools and techniques, it is possible
to present the information through multiple media. For example, we could build a master
database for the city to store all of the related data about land use. And based on the
database, an interactive platform is possible, allowing the users to select and present land
use information that interests them the most.
56
4 1 BIG DATA AND THE OPPORTUNITIES
In the history of land use practice, the capacity of applying a valid classification system
has been limited by the accessibility of information. With feasibility an important
principle to develop a land use typology, we will see an opportunity to extend the
practical scope of our framework, that is, the rise and application of big data.
4.1
The Concept of Big Data
It is hard to believe that there could be anyone who lives in the modern time has nothing
to do with the data. Not only our birth, marriage, sickness- the big things in our life are
recorded, but also our daily activities: everyday Facebook has more than 3 billion actions
of "Posting" and "Like"; google has to deal with 24 petabytes of data; twitter is doubling
the volume of data every year.....If we think about it, we store everything. And thanks to
the information and communication technologies (ICT) such as the internet and smart
phones, the data we are creating is in an explosive growth. Looking at our capacity to
create and store information (Figure 4-1), we could be like a man who suddenly gets rich
but does not know what to do. The question here is what should we do with these data
sets?
The answer could be revolutionary. In 2009, nature published an epidemiology paper by
researchers from Google. This paper explained how search engine query data could have
helped on estimating seasonal influenza epidemics. In the research, the engineers tracked
a series of influenza-related search queries with their online search engine, such as "what
are the medicines for curing cough." Then using past years' data, they built a connection
between some certain query entries and the data from the US Centers for Disease Control
and Prevention (CDC) and the European Influenza Surveillance Scheme (EISS). They
found a series of query combination for influenza detecting. And their estimation has a
97% correlation with the official data. More importantly, their detecting is immediate
while the CDC surveillance systems typically have a 1-2-week reporting lag (Ginsberg et
al., 2008).
57
The story brought a new perspective of data-that the current data collecting and
computing capacity makes it possible for us to keep track of the information for the
whole group instead of sampling them. As a result, large-scale and real-time modeling,
as known as "big data," is no longer a dream.
2W0r
ANALOG
18.86 billion gigabytes
Paper, flm, audiotape and vinyl: 6.2%
ANALOG
Analogvideotapes 93.8%
Other digital media: O.8%*
oIrA
Portable media players, flash drives: 2%
Portable hard diskm: 2.4%
CDs and mnidlsks: 88%
THE WORLD'S CAPACITY TO STORE INFORMATION
This chart shows the world's growth in storage capacity for both
analog data (books, newspapers, videotapes, etc.) and digital
(CDs, DVDs, computer hard drives, smartphone drives, etc.)
Computer servers and
mainframe hard disks: 8.9%
In gigabytes or estiUted equivalent
2000
Digital tape: 11.8%
1993
1986
ANALOG
2.62 MlNon
DVD/Bluray- 22.8%
ANA LOG STOR AGE
0.02 banOm
COMPUTING POWER
In 1986, pocket calculators accounted for much
of the world's data-processing power.
Pcentage of available processing powe by devlce
Personal
Pocket calculators
computers
1986
33%
2007
S6%
PC hard disks: 44.5%
1231111111a
Videogame Servers
consoles mainframes
9%
fgsbytes
17%
25%
30
Mobile phones.
PDAs
Supercomputers
0.3%
*Otheraiudeschrpcards, nvrsy
fokppy disks. mobile Phones/PDAs,
careras/am rerer., vides garies
ani,
2007
DIGITAL
276.12 billion gigabytes
Figure4-1 The World's Capacity Change to Store Information
Source: Researchersat the University of SouthernCaliforniatookfouryears -- 1986, 1993, 2000 and 2007 - andextrapolatedthe
numbers of roughly 1,100 sources of information.
Credit: ToddLindeman and Brian Vastag/ The Washington Post ("Rise of the digital informationage, "2011)
58
The concept of big data, in the beginning, means a collection of data sets so large and
complex that it becomes difficult to process them with traditional data processing
applications. But more and more, it became a phrase for a broad range of applications
that people could and only be able to achieve based on the large datasets (Mayer-
Schonberger & Cukier, 2013).
Big data is different with traditional data by its nature. The application of big data is
based on the complete collection of information rather than a sampling survey of the fact.
It focuses the correlations of the phenomena rather than asking the underlying causality.
In only a few years, big data has opened a new door for us to understand the world.
4.2
Opportunities for Urban and Land Use Studies
4.2.1
Broader Resolution and Scale
Source: Cheng, Caverlee, Lee, & Sui, n.d
ff
Figure 4-2 Scale And Resolution of Big Data
Scale is the "hierarchy" of spatial organization. It is considered as the fundamental of
geographic analysis. This is in accordance with the patterns of economic activities and
human settlement. When discussing the spatial scale of models, people use the term of
"resolution" to refer to the smallest geographic unit of analysis. And "extent" is a term to
describe the total geographic area of study (Agarwal, Green, Grove, Evans, & Schweik,
2002). Traditionally, when resolution becomes finer, the scale of the model becomes
smaller to keep the information at a manageable level. For example, if we look at the
global level, traditional models rarely include individual information such as location or a
single place.
59
Whereas the case in the era of "big data" is different. Despite the scale of study, the
resolution could maintain at a certain level, which means we could always keep the
information even when zooming out to the global level.
The advancement of mobile technologies has created an incredible amount of individuallevel data without executing a mass-scale survey (Figure 4-2). These data are of great
fineness of scale and have been increasingly used in regional, urban and population
studies (Long & Shen, 2013). Although the use of micro-models has been hindered at a
certain level by the poor availability of personal data due to privacy and cost constraints,
researchers have suggested methods such as agent-based modelling and micro simulation
could be complementary for individual-based modelling.
As Lam and Quattrochi have asserted, "Scale and resolution have long been key issues in
geography. The rapid development of analytical cartography, GIS, and remote sensing
(the mapping sciences) in the last decade has forced the issues of scale and resolution to
be treated formally and better defined" (Lam & Quattrochi, 1992). And now it should be
the time to redefine urban geographies as big data is enabling the possibility.
4.2.2
Dynamics Data Collection and Analysis
The city is a dynamic system. The traditional approaches of the city have overlooked or
oversimplified the dynamics of the city. This gap has a practical root. In traditional data
period, it is typically difficult and time consuming to collect and analyze data: think
about census data, huge amounts of investigators, almost one year long period, with an
average cost of 42 dollars per person. And it only happens every ten years. In confronted
of the ever changing society and the needs of the most updated data, the traditional way is
insufficient.
Compared with this, new approaches such as location based big data are providing us an
alternative to interpret urban activities. For example, with GPS facilities commonly
installed on buses and subway trains, it is easy to collect real time traffic data and make a
quick response to the changes (Figure 4-3).
60
Source: Senseable City Lab, MIT
Figure 4-3 Live Singapore, City Decisions in Sync
Source: Batty, 2008
Figure 4-4 Urban Complexity Study: PopulationMorphology andthe Road Network of London
4.2.3
Exploring the Complexity of Cities
Urban complexity has its basis in the regular ordering of size and shape across spatial
scales (Batty, 2008). By complexity I mean not only the self-similarity of urban structure
61
observed across the scales, but also the functional systems in the city (e.g. transit,
economic, social) that are highly interacted with each other. This matter of fact made any
effort of digitizing a city more like a single piece of puzzles. In this sense, the complex
systems study seeks simple non-linear coupling rules which lead to complex phenomena.
Complex system study (Batty, 2007, Figure 4-4), together with network science
(Newman, 2006) and big data in ascendant, is predicted to provide city a systematic
perspective to deal with its s complexity, although in the past, the focus on the city was
almost entirely on modeling traffic flows of the city (Wilson, 2010).
4.2.4 Urban Prediction
"Predictions based on correlations lie at the heart of big data" (Mayer-Sch6nberger &
Cukier, 2013), rather than a description of the fact. The capacity of big data on urban
prediction is based on its ability to analyze of correlations among vast amounts of data.
For example, Amazon could recommend us the ideal book, based on its huge data about
who buys what. It only needs to know people who bought the Lord of the Rings are more
likely to buy the Hitchhiker'sGuide to the Galaxy but not asking why this is the reason.
The growing respect for correlations of big data also brings a probability based method of
prediction. The reason is simple: with correlations, there is no certainty but only the
probability. Based on the historical data and correlation from a large group of samples,
algorithms will predict the likelihood that one conduct a certain activity. If the correlation
is strong, the likelihood of a link is high.
4.3
Land Use: Data and Beyond Data
The techniques of big data have been applied to demonstrate the patterns of human
activity such as mobility (Noulas, Scellato, Lambiotte, Pontil, & Mascolo, 2012), place
identity (Kottamasu, 2007), social media based civic movements (Sun, 2013) and public
participation (Schirra, 2013). Among them, there have been few researchers using mobile
phone data to classify urban land use (Pei, Sobolevsky, Ratti, Shaw, & Zhou, 2013;
Toole, Ulm, Bauer, & Gonzalez, 2012). Their researches are based on the assumption that
62
urban activities could be retrieved from the mobile phone data in order to indicate the
function of land use (Pei et al., 2013). These researches are among the first explorations
of big data on land use study although the concepts of land use and land use classification
in these researches are still in the boundary of conventions.
Why we still need categorization with big data--- Categorization is a short description
of fact behind the nominative implication. The real world is filled with so much
complexity and details that to fully describe them is neither necessary nor feasible. In
most cases, a group of key features could explain the most crucial characteristic of the
real world for our purpose of use. This is why in the theoretical study; we always need to
simplify the complex real world into a compact model. This model will be abbreviated
but could reflect the relevant truth that we need to know.
The term of "typology" is an aggregated description of the reality. This aggregation is
necessary because the amount of information is too large for us to perceive. Imagine
there are 13 dimensions of land use measurement (such as ownership type, the function
of the building, level of development etc.), how could we present the key nature of the
overall condition of a big area of land and make a comparison? Of course, we need
detailed information for narrowed-down examinations, but we also need a highly
structured, concise description of the reality.
Based on the involvement of big data, it is very possible that in the future, more and more
detailed data about human activity will not be a fantasy-which means we will be able to
collect information as detailed as the reality. For this reason, in this research, by
redefining the typology of land use, we are not only looking at breaking the limitations of
technologies and practices of land use, but also by typology, going back to the questions
on what are the key questions about urban land use.
Meet with big data, but go beyond big data. Big data as a revolution of technology
initiated the idea of this research. However, with the always improving capacity of data
gathering and processing, we have reason to believe in the future we will see more
revolutionary techniques. For this reason, we will not limit the scope of this research in
63
the boundary of big data. In contrast, we build the framework of the theory, even maybe
some of the techniques might not mature yet.
64
5 1 CASE STUDY: TESTING THE SNTAX WITH BIG
DATA
As a case study, this chapter will go through the process of developing a new land use
typology. I will pick two cities as examples, develop a big data based land use typology,
apply it, compare the new typology with conditional land use classification systems, and
evaluation the new typology that is built.
5.1
Example Cities
Two cities, Boston, U.S and Shenzhen, China are the example cities. The two cities are
distinct between each other on their development status and land use classification
systems that are used.
Table 5-1 U S Land Cover by Type (in million acres)
Year
Total surface
Rural land total
area
Developed
Water
Federal
land
areas
land
Land
2001
1,937.7
1,379.3
106.3
50.3
401.8
2002
1,937.7
1,378.1
107.3
50.4
401.9
2003
1,937.7
1,377.3
108.1
50.4
401.9
2001
100.0
71.18
5.5
2.6
20.7
2002
100.0
71.12
5.5
2.6
20.7
2003
100.0
71.08
5.6
2.6
20.7
Percent of total land
Source: StatisticalAbstract ofthe U.S 2012, Table 367. Land Cover/Use,
The city of Boston is in a context that about 70% of the total U.S land area are nonfederal
rural land (Figure 5-1). These lands are devoted to four primary categories: (a) cropland,
(b) pasture land, (c) rangeland and (4) forestland. In addition, about 40 million acres of
federal lands are designated as a wilderness area under the 1964 Wilderness Act. In total,
about three-quarters of the total U.S land area are considered as undeveloped. On the
65
other hand, about 10% of the land is devoted to urban uses, including developed lands
and federal lands.
The city of Boston is among the urban used land. Its land use information is collected and
organized statewide by the commonwealth of Massachusetts. The statewide data is
available for the year of 1971, 1985, 1999 and 2005. Land use was categorized into
21classes before 1999 and 37 classes in 2005. The original classes were more detailed
with up to 104 classes, but were aggregated into 21/37 categories.
The use of land in Boston is regulated by zoning code, which is granted by the Boston
Redevelopment Authority. Zoning regulates the uses, dimensional boundaries and height
of privately owned buildings and land. The current City of Boston Zoning Code was
enacted in 1964 and has been modified numerous times. There are fourteen
neighborhood codes plus eighteen codes for downtown and the waterfront.
Boston is one of the oldest U.S cities that its documented history could trace back to the
early settlement in 1630 by Puritan colonists from England. The city has been on the
forefront about social and urban development, including the first subway system in the
U.S. The city has gone through its development period with most of its available land that
is developed. In terms of urban management, the city and the state have been
implementing the Open Government & Data strategy, and now most of the published
official data could be easily reached.
Table 5-2 ChinaLand Cover by Type (in ten thousandshectares)
Year
Cultivated Land
Forests
Water Area in Land
Area of Grassland
2004
13004
17491
1747
40000
18.22
1.82
41.67
-
Usable Area
Others
23758
-31333
Percentage of total land
2004
13.55
--
32.64
24.75
Source: StatisticalAbstract of China Statistics2005
On the other hand, Shenzhen lies in a country where the construction land has been
growing rapidly (Figure 5-2). Shenzhen is one of the youngest cities in the world. It was
born with China's Reform and Opening Up policy. In the year of 2010, the city's total
66
population reached 10 million, while in 1979 when Shenzhen was established as China's
first special economic zone the population was only about 300 thousand-Shenzhen is
one of the fastest-growing cities in the history of mankind, and it is still growing at a
certain speed. Under this condition, it is hard to imagine that the conventional way of
measuring land use could fulfill the need of fast-paced urban change and management. In
addition, although China has made a move to open government data in recent years,
getting official land use data is still extremely difficult.
In overall, the two example cities provide different contexts for developing the big data
based land use typology (Figure 5-3).
Table 5-3 Quick Fact of the Two Example Cities
Shenzhen, China
Boston, US
Major city in Guangdong
Capital and the largest city of the
Province, China
Massachusetts State, U.S
Year of incorporationas a city
1979
1822
Growingspeed
Very fast
Slow
Official data accessibility
Low
High
Urban activity
Active
Active
Position
Generated by the author
5.2
Big Data Mining and Process
5.2.1
Main Datasets
Among the various big data sets, I choose geosocial network data as the main data set.
Specifically, I will use data from Foursquare.com for the city of Boston and its Chinese
counterpart Weibo.com for the city of Shenzhen.
Geosocial networking is a type of social networking based on geographic services. The
geosocial network services allow the users to interact with people to their current
locations. Compared with other forms of social media, geosocial networking can be used
with geotagged information to match users and their activities with a place, event or local
67
activities. Most of the popular social media now have geosocial sections, such as
Facebook Places, Twitter Location Feature.
Among them, Foursquare is a website which is mainly designed for location based
networking on mobile devices. Foursquare allows users to "check in" at nearby venues
using a mobile website, text message or smartphone apps. The location of the user is
based on the GPS section of the mobile devices or network location provided by the
operators, and the maps are based on data from Open Street Map, which is an opensource map provider. Similar to social media websites, Foursquare allows registered
users to post their location and connect with friends. Foursquare started out in 2009 and
rapidly reached 7 million users in 2011. Compared with Facebook, Twitter and other
most popular social networking services, Foursquare has a smaller size of registered
users, but its services are 100% based on geotagged information, which is ideal for our
spatial analysis.
As a counterpart in China, Sina Weibo is a combination of Twitter and Facebook. It was
launched in 2009, the same year as Foursquare, and has about 503 million registered
users by the end of 2012. "Weibo" is the Chinese word for "microblog." And although
there are other Chinese microblogging services such as Tencent Weibo and Sohu Weibo,
the term "Weibo" sometimes is directly referred to Sina Weibo because of the fact that it
is one of the most popular social media in use by 30% of Internet users. I picked Sina
Weibo rather than Jiepang (the Chinese version of Foursquare), as our main data source,
because Weibo is a service covering a wider range of user groups.
As indicated in the previous chapter, big data is a broad topic that is covering more and
more aspects of our life. Social media is a growing section and a representative of this
trend, where most of the content is generated by the users rather than the operators.
Geosocial media, specifically Foursquare and Weibo here, provide us a perspective on
the city's land use from the user's point of view, enabling us to know how the urban
spaces are used and evolved in the user's daily life. For example, while we were
struggling to assign the place with a function, the data from social media provides us the
68
user identified function of the venue. The data are interactive and dynamic, from what we
can track the intensity and change of land use through time. In this sense, the urban
places are what they are in daily use, restaurants, events, offices, residences. The
reliability of the data is based on the collective perception from a big amount of people.
When comparing the two examples of Boston and Shenzhen, it is important to keep the
fact in mind that the two platforms of Weibo and Foursquare are different. One of the
differences is the overall development status of mobile phone based social media in
China and in this U. S-smart phones and smart phone based social media services are
more widely developed in the U.S. Another difference is the recognized function
Foursquare and Weibo: while Foursquare is designed as a location-based social media
platform, web is more recognized as an online community, whose venue database is not
the core of its use.
5.2.2
The Techniques to Acquire Data
To acquire the social data about land use, we employed application programming
interface (API) through which social media platforms share content and data and
encourage third party developers to develop web and mobile applications.
Both Foursquaer and Sina Weibo provides the certified developers with API services.
And they both have its special section for venue information. The venue databases allow
users to search and find information about venues and places, such as the number of
people who have "check in" at a place, the tips or the photos about the place.
69
Figure 5-1 Gridfor DataCollection
Generatedby the author,base mapfrom Open Street Map
Searches can be done with by locations or by keywords through the city. The basic input
parameters of searching venues include the latitude and longitude of a location, the
authorization number of the user, plus other optional parameters to define the way of
searching. The output of the searches includes compact information about venues nearby
the input location. The information often includes the venue's ID, name, its accurate
location, its address, the category that it belongs to, the number of check-in, etc.
To acquire the venue data of the cities, I use a grid system to search the nearby venues'.
The data were collected and parsed with Processing, a Java based programming language
and development platform.
1 The Foursquare API permission is from the research project of We Are Here Now, supervised by
Professor Sarah E. Williams from Civic Data Design Lab at MIT. The Weibo APIs accounts are from the
Creative City project, cooperated between Civic Data Design Lab and Urbanus, Hong Kong.
70
5.3
Composing the Typology
In this part, I will follow the principles and process that are defined in the syntax, and
develop a social media based land use typology with Foursquare data and Weibo data for
Boston and Shenzhen.
5.3.1
Topic of Interest/Purpose
The purpose of the typology is to examine urban activities and land use patterns in the
context of big data. Different from the conventional land use survey process, the data is
generated by the crowd rather than the professional planners. With that, we can expect an
alternative to interpret urban land use from the resident's point of view.
5.3.2
Defined Scale and Unit of Measure
The scale of the case study is, spatially at the city level, and timely daily based. As for
the unit of measure, object based unit, such as land use parcels, blocks or buildings, will
be the ideal option. If not applicable, the grid based unit is acceptable.
5.3.3
Defined Elements
In this case, land use is indicated by social media activities in the city. Specifically, I
define land use function as the category of venue that is assigned by the website, which
is not necessarily a scientific classification but a recognized identification of the urban
venues. The intensity of land use is measured as the number of check-ins at a place. The
measure might be biased, but it is an indicator of urban activities from the perspective of
social media. Accordingly, land use connectivity is interpreted as the level of social
connection, that is, in this case, the similarity of visitors to the place. Land use
probability here is not directly mentioned, but applied through the discussion of land
mixed use. It mainly refers to the proportion and combination of land use functions in the
mixed used land.
71
5.4
BOSTON Result
5.4.1
Data Briefing
Table 5-4 Boston, the Input of Query
Field
LL
Limit
Intent
SW
NE
Description
Latitude and longitude of the query search
Number of results to return, up to 50.
Indicating the intent in performing the search. If no value is specified, defaults
to check-in
Limit results to the bounding quadrangle defined by the latitude and longitude
given by seeing as its south-west corner, and NE as its northeast corner. Not valid
with LL or radius. Bounding quadrangles with an area up to approximately 10,000
square kilometers are supported.
See SW
Used
50
Defaults
Source: https://developer.foursquare.com/docs/venues/search
As described in the previous part, I use Foursquare API -search venues to collect the
desired land use data. Foursquare allows users to identify the parameters of the query of
data. The parameters used in this research are listed in Table 5-4.
The response of query includes an array of compact venues, each of which contains a list
of contents. In this case, I only stored the related fields of our interest (Figure 5-5).
Table 5-5 Boston, the Output of Query
Field
1In
Description
Response
L
Id
A unique string identifier for this venue.
The best known name for this venue.
An object containing none, some, or all of address (street address),
cross street, city, state, postal code, country, lat, Ing, and distance.
Location
All fields are strings, except for lat, Ing, and distance. Distance is
measured in meters.
An array of categories that have been applied to this venue. One of
the categories will have a field primary indicating that it is the
.
Categories
primary category for the venue. For the complete set of categories,
see venues/categories.
Contains check-ins count (total check-ins ever here), users count
Stats
(total users who have ever checked in here), and tip count (number of
tips here).
Source: https://developer.foursquare.com/docs/responses/venue
Name
72
I/
Used
For Boston, The data collection program was tested from February 21 to May 4 2014,
and operated during March 8h through March
2 1s",
2014. For each day the program was
started at 10:00 PM. It takes the program 4-5 hours to complete the query. The collected
data are points that cover the whole area of Boston city with an acceptable level of
density (Figure 5-2).
Figure 5-2 Boston, the Output of Data
Generated by the author
Besides of the venue data, other datasets and maps are collected:
a. Boston city boundary (shapefile)
Source: The City of Boston
Description: Polygon of the entire city of Boston
b. City of Boston parcel data 2013 (shapefile)
Source: The City of Boston
Description: This polygon layer contains all of the property parcels in Boston.
Each parcel has a shape and a unique number that links it to the record in the
73
Assessing Department's main Parcel Inventory system (OWNHIST). The data
file has no access and use limitations.
c.
City of Boston building footprint data (shapefile)
Source: The City of Boston
Description: the City of Boston building footprint data. Created in 2012 by
conflating three data sources. The 2011 planimetric buildings, SAM (Street
and Address Management) buildings, and BRA buildings. The fields included
in this dataset are listed in Table 5-6.
d. Massachusetts street transportation dataset (shapefile)
Source: Office of Geographic Information (MassGIS) - Mass. Gov
Description: official state-maintained street transportation dataset with local
and major roadways, including designations for Interstate, U.S. and State
highways. The layer is up-to-date through December 2007.
e. Boston land use data (shapefile)
Source: Office of Geographic Information (MassGIS) - Mass. Gov
Description: a statewide, seamless digital dataset of land use / land cover for
the State of Massachusetts derived using semi-automated methods and based
on digital images captured in 2005 with 0.5 m pixel resolution. The minimum
mapping unit (MMU) for this dataset is 1 acre overall for the dataset. The land
use classification scheme used for these data is based on a coding scheme
used with 37 land use classifications.
parcels_13.shp
idar.alLshp
d EOTROADS ARC sh
Figure 5-3 Boston, Parcels,Buildings andRoads
Source: The City of Boston and Commonwealth of Massachusetts
74
Table 5-6 Boston, BuildingFootprintData Fields
FIELDS
DESCRIPTION
BASEELEVATION
Elevation of structure (NAVD88) one decimal place
ELEVATIONSEA LEVEL.
decimal
Elevation of roofline edge above sea level (NAVD88) one
place
Elevation of the highest point above sea level (NAVD88) one
TOlPSEALEVEL
decimal place
ELEVATION_GROUND LE
Elevation of roofline edge above ground level (NAVD88) one
VEL
decimal place
Elevation of highest point above ground level (NAVD88) one
TOP GiROUIND LEVEL
decimal place
Source: City of Boston
5.4.2
Unit of Measure
The case study applies to the entire city of Boston. Based on the available data sources, I
use parcel and building footprint as the basic unit of measure (Figure 5-3), where
Foursquare venue data will be aggregated into. Yet the resolution of venue data allows us
to use grid or other spatial unit system, our intent here is to test the theory with this
specific unit system. In Shenzhen case, I will apply a grid system as the unit of measure.
5.4.3
Land Use Elements
A. Function
Through Foursquare data, each venue is categorized according to its usage, such as bars,
parks, department store or Thai restaurant. These detailed categories are organized into a
hierarchy containing primary categories, sub-categories and sometimes sub-subcategories. For the case study, I use the primary categories as the indicator of land use
function. They are: food, nightlife, outdoor & recreation, residence, shop & service,
travel & transport, art & entertainment, college & universities, event, and professional &
others. While the land use analysis will mainly be based on the 10 primary categories, in
this case, when putting into other implementations, it is feasible to use more detailed
categories of venues if needed.
75
t
Gnrte
J(
0
Venue Types
%Ji
Figure 5-4 Boston, Categories of Function
Generated by the author
The figure 5-4 shows the 10 categories and their subcategories. From it, we can see Boston
has its most detailed categories among food and services. The following chart, Figure 5-5
shows the count of parcels by each type of function, by which we see that Boston has its
biggest share of venue counts on residence. Professional & other places counts for the
second. These two types of venue represent the basic functions of home and work in the
city. Other function such as transportation, recreation, shopping, education, shares a
smaller amount.
76
(BLANK)
TRAVEL & TRANSPORT
SHOP & SERVICE
RESIDENCE
PROFESSIONAL & OTHER PLACES
-O
OUTDOORS & RECREATION
NIGHTLIFE SPOT
FOOD
EVENT
'
COLLEGE & UNIVERSITY
-.-
ARTS & ENTERTAINMENT
Figure 5-5 Boston, the Count of Parcelsby Function Types
Generatedby the author
B. Intensity
The figures below are maps with the intensity of land use measured by building height
(Figure 5-6-a), the total number of venues in the building (Figure 5-6-b), the total checkin number of a building (Figure 5-6-c), and daily check-in number of a building (Figure
5-6-d).
The result shows significant variations. By building height, the most intense areas of
Boston are through downtown and the prudential area, and expanded to Allston.
Compared with that, the distribution of venue intensity and check-in intensity are more
extreme, concentrating in transportation hubs (e.g. the Logan airport, South Station,
North Station) and commercial centers (e.g. Prudential center, Back Bay).
The differences here emphasize various aspects of land use-the building height
indicates the built environment, which is a reflection of the rental price of land; venue
intensity could indicate a commercial and recreational center of the city, where
concentrated services and shops are beneficial; transportation hubs are places of
interchange and movement, therefore most checked by people.
77
(a)Building height ofBoston: high buildings distributein the CBD area and the commercialcenter, in
accordancewith rentalprice of land (b) Venue intensity ofbuildings: the financialdistrict is less intense
than commercialand recreationarea.
78
(c) Total check-in intensity of buildings: dense area includes the airport,South station, North Station,
Prudentialand Boston University. (d) Daily check-in Intensity of buildings: the transportationhubs are
significantly higher than other areas.
Figure5-6 Boston, the Intensity of Land Use
Generatedby the author
79
5.4.4 Generic Applications
A. Mix Use
Single use 100%
- .
*Dominate
use >50%
Mix use
Figure 5-7 Boston, Mix Use of the City
Generatedby the author
As is discussed in chapter 3, mix use mainly focuses on function and the probability of
function. To illustrate mix use of land parcels, I divide Boston land parcels into three
groups: single use, dominate use and mix use (Figure 5-7-Figure 5-10). The
classification is based on the functional proportion of venues located in the parcels.
Among the 10 primary categories of functional venues, if a single category counts 100%
of the venues in a parcel, the parcel is defined as single used; if a single category counts
more than 50% of the venues in a parcel, the parcel is defined as dominate used; if none
80
of the primary categories counts up to 50% of the venues in a parcel, the parcel is defined
as mix used.
Proauinal
Other
shop
.~4
&serv.c
,
ghaliespot
A--
U,4F
Colege & Univenes
Travel $ Tansport
*
/
.1
Figure 5-8 Boston, Single Used Parcelsby Function
Generatedby the author
81
$
a
~
I
*
.,*~ %
0
7
vwd
Fe
7
$
N
A4
Mk$
Figure5-9 Boston, Dominated Used Parcelsby the DominatedFunction of Use
4,1
~4
-Pu-
N
V/
4'
Shop
N
1
o~~
1'us
Figure 5-10 Boston, Mix Used Parcelsby the Most Frequent Type ofFunction
82
Diveruityl
N
Figure5-11 Boston, Land Use Diversityas an Index of Mix Use
Generatedby the author. Simpson' diversity index
S
=
, where Pi is each type's proportion
among the total number
To measure the level of mix use, I borrow the diversity index from the school of ecology.
The Simpson Index, specifically, is powerful to measure the degree of decentration when
individuals are classified into types. It represents the probability that two individuals
randomly selected from a sample belong to different species. The value of the index
ranges from 0 to 1. The bigger the value is, the higher the diversity is. I calculate
Simpson diversity index for each parcel, which is illustrated in the map (Figure 5-11).
83
We could understand the diversity index as a comprehensive measure of mix use, where
0 means singly used parcels and 1 means fully mix used parcels where each venue
belongs to different functional groups of land use. It is also worth mentioning that since
the venues are classified only into 10 primary categories, even in the defined single used
parcels if looking at the sub-categories of function, the parcel is not completely
homogeneous.
B. Land Use Change
ParcelLand Use Change
-926.0
-9259 --245.0
-244.9 - -40.0
-39.9 - -5.0
-4.9 - 5.0
N
5.1
4 Mies
1
-
12.0
12.1 - 43.0
Figure 5-12 Boston, Land Use Change in a Week
Generated by the author
To demonstrate the change of land use, I picked a typical Sunday (March 2 nd) and a
typical weekday (February
2 7 th)
for the comparison. The result is shown in Figure 5-12.
From the map, the land use of the city has a minor change due to the short period of
84
observation. The most significant land use change happens at the airport (green area in
the northeast corner). This means there are much less traveling people on Sunday than
weekdays.
Due to the limitations of data collection, land use change is only conducted in a weekly
framework of time. If individual check-in data or data for a longer period of time is
available, we could expect to see land use change of the city by the hour, by month or by
year.
5.5.1
Types of Land Use
0 Art
* Event
0 Food
Nightlife * Outdoor
* Profession
* Residence * Shop * Travel * University
5.
50,000
*
40,000)
3!"000
0
S
%s*
0
0
30,000
00
0
I
0,
**
a 0
0*s*
#
se
0
04
*
20.000
0
0
0.00
it:
0.10
15,000
00
so
10,000
010
0.30
0,40
0.50
0.60
0.70
Single Use
0080
M90
Mix Use
5
Figure 5-13 Boston, Land Use Types
Generatedby the author
In the case of Boston, the typology of land use is examined with land use function, land
use intensity, land mix use and land use change. With the variances, we could classify the
Boston land use parcels with three dimensions: land use types, condition of mixed use
85
and land use intensity. Here land use change is not included as a variance because its
variance among the parcels is minor.
The figure 5-13 shows one of the possible classification of the land use, where the parcels
are shaded according to function, and are plotted according to its level of mix use and
intensity of use. From the figure, we can identify four significant groups of land use
parcels. The group I contain single used parcels where the intensity and function of land
use vary from each other. Group II includes parcels that are not fully mixed used nor
intensively used. Group III are parcels in mixed use, but with a relatively low level of
intensity. Group IV are parcels that are both mix used and with a high intensity.
The groups only present a visual classification of the land use parcels in Boston. In the
practical use of the method, we could include more dimensions into consideration and use
methods such as statistics to classify the parcels of land.
Source: the commonwealth ofMassachusetts
9RUSHLANDISUCCESSMAL
OPEN LAM
*TL"
K4 MN-FORESTED WETLND
L SALT WATER WTLAND
SFORESTE
SALTWATER SANDY BEACH
C
C
DNRuOG
CROP LAWt
PASTIRE
M
GOLF COURSE
PARICPATION REC
SSPECTATOR REC.
LICH
OMBTY RES.
MMHKJ DEMTYRES
LOW DEMITY RES
VERY LOW DENMIT
RES.
TR"1T1D1NfL
URBAN PWLIC/M6TITUTIONAL
COMMERWA>
MTRANPORTATFM
SPOWEOUNEUTIUTY
Figure 5-14 Boston, Existing Land Use Typology
Compared with the current land use data (Figure 5-14), we have expanded the dimensions
of land use from a single label to a comprehensive interpretation of function, intensity,
86
mix use and land use change. The process of determining land use classes is clear and
tells information about land use patterns of the city.
5.5
SHENZHEN Result
5.5.1
Data Briefing
Similar to Foursquare for Boston, Weibo API allows users to specify the parameters of
the query. The parameters used for Shenzhen are as below:
Table 5-7 Shenzhen, the Query of Data
Field
Description
Used
Lat
Latitude of the query search
Long
Longitude of the query search
Range
Query radius. The default is 2000, the maximum is 10000, by meter
Count
Number of results to return for each page, up to 50. The default is 20; the maximum is
50.
Page
The page number if there is more than one page of results to return. The default is I
300
50
Source: http://open.weibo.com/wiki/2/location/pois/search/bygeo/en
The response of the query includes an array of compact venues, each of which contains a
list of contents. In this case, I only stored the related fields of our interest:
Table 5-8 Shenzhen, the Output of Data
In
Response
Field
Description
Poiid
A unique string identifier for this venue.
Title
The best known name for this venue.
Location
An object containing none, some, or all of address (street
address), city, province, postal Code, country, lat, Ing, and distance.
An array of categories that have been applied to this venue.
Contains check-in num (total check-ins ever
here), checkin user num (total users who have ever checked in
Statshere), and tip num (number of tips here), photo number (number of
photos here).
Source: http:/open.weibo.com/wiki/2/location/pois/search/by_geo/en
Categories
87
Ussed
The data collection program was tested for Shenzhen from March 9 to May 13 2014, and
operated during March 15t through March 21s, 2014. For each day the program was
started at 12:00 AM. It takes the program 5 hours to complete the query.
Figure 5-15 Shenzhen, the Output ofData
Generated by the author
Other datasets collected for Shenzhen include:
a. Shenzhen events data
Source: Douban API
Description: Douban is a website where people can record, comment, and create
content for films, books, music and events & activities in Chinese cities. Some
individuals (e.g. authors, artist), groups (e.g. indie bands, musicians) and
organizations register their pages on this site. And they are often the creators of big
events. The data are all relocated Douban events for Shenzhen during June 1 s-7th
with event details, number of participants and the ID list of participants.
b. Shenzhen base maps (shapefile)
Source: Open Street Map data.
88
Description: open source data about roads, trails, caf6s, railway stations and
locations. The data are not ensured to be complete and most updated.
5.5.2
Unit of Measure
Lacking parcel and building footprint data, the Shenzhen case uses grid based system as
the unit of measure. The size of the grids will vary as needed.
5.5.3
Land Use Elements
A. Function
For Shenzhen, the land use function is again defined by venue categories assigned by
Sina Weibo. There are more than 200 parallel categories from Sina Weibo, which are not
organized as a hierarchy. For easy comparison with the Boston case, the categories are
translated and assigned with the primary categories that are the same as Foursquare.
There are 9 primary categories assigned to all of the Shenzhen venues (Figure 5-16).
They are: food, nightlife, outdoor & recreation, residence, shop & service, travel &
transport, art & entertainment, college & universities, and professional & others. While
Food is again a significant category, the category of "Event" in Boston case does not
show up in Shenzhen case.
89
o~Port/
Venue Types
Figure 5-16 Shenzhen, Categories ofFunction
Generatedby
the author
Figure 5-17 shows the distribution of Shenzhen land use cells among the 9 categories.
Compared with Boston, Shenzhen demonstrates a different social activity pattern:
Professional places and food are the top two places of Weibo venues. The followed are
commercial place, residence, and transit stations. Recreation and educations places, such
as nightlife, universities in Shenzhen are not as significant as the Boston case.
90
TRAVEL & TRANSPORT
3115
SHOP & SERVICE
4351
RESIDENCE
3835
PROFESSIONAL & OTHER PLACES
7251
M
OUTDOORS & RECREATION
NIGHTLIFE
1177
529
FOOD
COLLEGE & UNIVERSITY
7197
M
253
E
ARTS & ENTERTAINMENT
0
804
1000
2000
3000
4000
5000
6000
7000
8000
Figure 5-17 Shenzhen, the Count of Cells by Function Types
Generatedby the author
B. Intensity
While intensity could be interpreted in different ways, in this case of Shenzhen, I use
Kernel Density to demonstrate the intensity pattern of land use. The idea of Kernel
Density is to calculate a magnitude per unit area from point features using a kernel
function to fit a smoothly tapered surface to each point. In other words, each point values
the density of features in an area around the point. In ArcGIS, the output of Kernel
Density is based on raster cells. Conceptually, a smooth curved surface is fitted over each
cell. The surface value is highest at the location of the point and diminishes with
increasing distance from the point, reaching zero at the Search radius distance from the
point. The volume under the surface equals the Population field value of the point, or 1 if
NONE is specified. The density at each output raster cell is calculated by adding the
values of all the kernel surfaces where they overlay the raster cell center.
91
X*
~Ae
A
u"Cont
4
~ -
s
022545
n 22OW
)N
i
~
A enI
-
ep
i
ml
Im
Lo W
0
Coowtin
225 4.5
s
--
1w
r
Ccki
-syPOWy
4.
2
-;
2.25
i I 1+-,mAn
Figure 5-18 Shenzhen, the Intensity of Land Use
Kernel Density for (a) Venues, (b) Total Check-in, and (c) Daily Check-in
Generatedby the author
92
The Kernel Density maps of Shenzhen show the distribution patterns of land use intensity
of the city: the density of places (Figure 5-18-a), demonstrated by Sina Weibo venues, is
higher in the central area and development areas than in small towns and mountainous
areas. It is also the case for the density of activities, demonstrated by Sina Weibo venue
check-ins (Figure 5-18-b and Figure 5-18-c). In general, the activity density is more
spatially concentrated than the venue density.
C. Connectivity
In this case, I use Douban event data as an indicator of land use connectivity. The level of
connectivity between two events is defined as the number of people that went to both of
the events. Figure 5-19 is an illustration of the concept, where each circle represents an
event. The weight of the lines represents the number of participants that two events have
in common, and the size of the circles represents the popularity of the events.
4
4-
Figure 5-19 Shenzhen, the Connectivity of Land Use
Generatedby the author
93
The connectivity of land use here could be understood as part of the social network of the
city. It could help to estimate the transportation flows of the city. And in the future, the
same concept could be applied to land use-transportation integration if data is available.
5.5.4 Land Use Application
A. Mix Use
The analysis of mix use is based on a 500 meter grid system covering the whole
municipal area of Shenzhen. Again, I define the level of mix use, according to the
probabilities of functional venue types in each cell. The 9 primary categories of venues,
which are defined in land use function part, are calculated proportionally for each cell of
the grid. If a cell is 100% occupied by a single category of venues, it is defined as single
use. If a cell has more than 50% of venues belonging to one category, it is defined as
dominate use. If there is no category in a cell counting for more than 50% of the venues,
the cell is defined as mix used. The results are shown in Figure 5-20.
B. Intensity and mix use
The new typology of land use allows not only a combined focus on the land use elements,
but also a combination of selected elements and packages. As an example, this case
combines land use intensity and mix use. With a 500 meter grid, I calculated the diversity
index for each cell and created a smooth surface with the function of Inverse Distance
Weighted Interpolation in ARCGIS. For the intensity, I added up the number of venues in
the same column of the grids, and created a graph to show the horizontal distribution of
the venues.
94
J
,a
a
e
do
goo
.0
.
1* *
.
-
as
--
a
a
L1
3T.
5ky
a
we
*s
0
2.5
5
1
us.
v-.
-
.
m
- .
-
Fu 5
.e
a
-
*If
9.
)
1
-
Us
tosaO
Geere
basa
a Aa
r0w .0
OR
a
a.
Ge er te
9
~
r.
by 95the1au ho
k..IA.
8s
The result provides a comprehensive illustration of the city's land use pattern, where we
could see the level of mix use of the whole city and aggregated information of land use
intensity among different functions (Figure 5-21).
As a matter of fact, the volume of information that could be illustrated by 2D maps is
limited. Therefore, I propose to combine the land use typology with an interactive
platform, where the user could pick their interested area and customize the way of
illustrating the data.
. Travel & Transport
t Food
" Shop & Service
" College & Universities
a Professional & Others
" Residence
Outdoor & Recreation
a Nightlife Spot
eArt & Entertainment
Hg:08633406
0
2.25 4.5
9miles
Low: 0
Figure5-21 Shenzhen, Land Use Intensity And Mix Use
S
Generated by the author. Diversity index is calculatedwith the Simpson diversity index
b -1 -jZn'
i-1
; Pi is
each type's proportionamong the total numbers; intensity is calculatedas the total count ofvenues along
the same rows of the grid.
96
5.5.5
Types of Land Use
500
con
400I
e! 3Mix
Use High Connecivty
&200
Figure 5-22 Shenzhen, Land Use Types
Generatedby the author
The land use of Shenzhen is examined with three elements-function, intensity and
connectivity, mix use, and a combination of mix use and intensity. With that, I would like
to classify the land uses with three dimensions of land use intensity, connectivity and
diversity. In the graphic above (Figure 5-22), each point represents a cell of the 500meter grid system. For each cell, I use Simpson diversity index is as an indicator of mix
use, the logarithm of total check-in numbers as an indicator of land use intensity, and the
total number of event participants as an indicator of land use connectivity.
Based on the statistical distribution of the points. I identify four groups of land use cells.
The first two groups are low connectivity groups, with one group of single used cells and
another group of mix used cells. The third group included land use cells with low level of
97
mix use and high level of connectivity. And the fourth group are land use cells that are
both mix used and well connected.
Again, these land use groups might not be statistically significant nor taxonomically
complete. It only presents a visualized classification of land use with accessible
information.
Source: Beijing MunicipalLand Use Planning(2006-2020)
N
SCULTNATED LAND
,
0AGARDEN
LAND
FOREST
GRASSLAND
OTHER AGRICULTURE
URBAN LAND
RURAL RESIDENTIAL
MINING LAND
SPECIAL USE
TRANSIT FACILITIES
0 5
UNUSED LAND
102030
K
WATER BODY
Figure5-23 China,Existing Land Use Categories
As a summary, first, the social media based land use typology provides an accessible data
source for Shenzhen, where the official land use data is not open to the public. Second,
compared with the current land use classification systems (Figure 5-23), we learn about
the land mixed use, the intensity and social connection of land use from land use
typology developed here.
98
5.6
Reflections
In the case study, I have used social media data as an example to apply the syntax of land
use typology. I defined land use function with social media activity types, land use
intensity as the strength of activities, land use probability as the proportion of different
activities, and connectivity as the bound of social network. I also defined the unit of
measure and scales based on the quality of data and capacity of data processing.
Yet we should acknowledge that the definitions used in the case study are not the only
ways to interpret the concepts of land use elements and that social media data should not
be the only data source for implementing a new typology of land use.
5.6.1
Limitations
The case studies have limitations in terms of data source and the method. First of all, I
use social media data as the main source of big data. However, social media data has its
limits and bias:
a) The users of social media are specific demographic groups of people. The users of
social media services might be more young people than old people, more urban
residents than rural villagers. Some of the services, such as location tag or checkin activity, are available only to those who has a smart phone.
b) Social media activities might not fully represent the actual urban activities. For
example, people tend to check in at places that they do not regularly visit; the
most checked places might be tourist places, landmarks or city attractions, which
are not daily places; the places where people check in are destination rather than
half-way places, etc.
c)
The quality of social media data might vary across regions. In the case studies,
Foursquare data for Boston has better quality than Weibo data for Shenzhen, not
only because Boston has more percentage of mobile based social media users, but
also because the Foursqure platform is more developed and specifically designed
99
for location based services. The difference between social media platforms makes
the results not comparable between each other.
Second, data are collected with a rough resolution and in a short period of time, making it
unfeasible to illustrate the ideas developed in the theory part. For example, as the land
use connectivity analysis for Shenzhen only includes events that happened through June
1st to June 7th, 2012, the overall connectivity might be underestimated. For Boston, as
data are only collected every 24 hours for one week, it is only possible to compare land
use change by weekday and weekends, although land use change could also happen by
hours of the day or by a long period of time. If we are able to acquire individual check-in
data, we might be able to examine more phenomenon such as mix use by day and night or
land use change by minutes.
Moreover, the proposed concepts and generic applications of the syntax are not fully
tested in this case study, because of my limited capacity.
5.6.2
Evaluation
Table 5-9 Evaluation ofthe SocialMedia Based Land Use Typology
tUnecK-m numDer as an inucaior 01 intensity is not
Depends on the data
accurate, while information about venues are
selected
relatively reliable
Both check-in numbers and venue information might
be biased
Activity based land
Activity based land use, mix
Multiple dimensions, an
use, mix use,
use, land use connectivity, a
open system that was not
dynamic observation
comprehensive interpretation
enabled by conventional
of land use change
of intensity and mix use
land use typology
Operable; measures are open to be developed
The method is adaptable to other data sources; the
categories used could be associated with
conventional land use classes
100
High
If compared with the conventional land use classification, the social media based land use
typology developed and tested in this chapter is significant in five aspects:
a) It focuses on the content of interest, in this case, the activity aspect of land use,
while the physical features and ownership of land, considered as external
variables, are not discussed.
b) The new typology has a rich meaning, with more dimensions of land use included
in the measurement. By that we can not only know the types of land use, but also
be able to retrieve information about mix use, the intensity and the overall land
use patterns of the city.
c) While the conventional land use classifications provide instructions to classify the
use of land, the process of big data based typology is flexible, where the elements
could be adapted according to the data source, the need and interest of study.
d) The data collection process is less time and labor consuming than the traditional
land use survey process. Therefore, the land use information could be instant and
dynamic. If with the full access to the data source, we could even expect a
dynamic land use observation.
e) The typology also reserves space to connect with the conventional land use data,
which the defined functions and categories could be assigned to conventional land
use types.
Currently, the new system could not replace the conventional land use classifications.
There are four reasons. First, the scope of social media based land use typology is on
developed areas, it will not cover land use types without human activities such as the
water body and wild land. Secondly, the system is flexible and open. Its development is
highly dependent on what the data source is, and what measures to use. It will be difficult
to form a standard with such a system. Thirdly, the theory framework of the new
typology is much more complex than the conventional land use classification, which
could confuse the untrained users. Last, current big data has its own problems, therefore
in practice, the reliability and validity of the new typology are not ensured.
101
5.6.3
Implementation and Impact
The social media land use typology could be used as a complement of the traditional land
use information. It is of interest to planners and the administrators of the city. On one
hand, this typology provides a different perspective of land use, by which the residents'
activities are emphasized. From that, the planners could observe urban land use beyond
the ownership, the built environment and physical features of the land, therefore better
propose for human-respect planning.
On the other hand, the social media based data is instant and quick collected, which is
sometimes more important that the accuracy of data. With the tools of social media data,
the planners could keep a constant track of urban land use, and respond quickly to any
unusual activities that are observed.
Moreover, this typology developed the descriptive measures for some of the land use
phenomenon that have never been measured before. For example, as land mixed use has
been proposed as a planning concept, there was no widely recognized method to quantify
the level of mix use. The diversity index in this case study provides an alternative to
make mix use discussable and comparable.
102
6 | CONCLUSION
Generatedby the author
THE PACKAGES
THE ELEMENTS
THE APPLICATIONS
Land Use Typology
for Energy Analysis
Function
Land
Mixed Use
Land Use Typology
for Urban Planning
Intensity
Land Use
Conflict
Land Use Typology
Probability
Land Use
for Land Resources
Land Use Typology
for Transportation
Change
Connectivity
Land Use
Estimation
Scale........|
Figure 6-1 the Syntax of Land Use Typology
By reviewing current land use classification systems, I summarized the problems of
current land use practices and posed the question of what a valid land use typology
should look like. To answer this question, this thesis did not propose an actual land use
classification system, but proposed a framework in terms of how to build a land use
typology. That is the syntax of land use typology (Figure 6-1).
The syntax includes five essential elements of land use typology: function, intensity,
connectivity, probability and scale. The thesis proposes that the land use typology
system is a package where the elements are defined specifically by the context, such as
energy analysis, city planning, land resource management and transportation study. The
concept is that the elements of land use typology could be defined with different
measures. With different purpose of use, the developers of land use typologies should be
able to pick and develop the land use elements with the most relevant measures for their
needs.
103
With that as a framework to build a land use typology, the developed land use typologies
could be applied in the study of mixed use, land use conflict and land use change. The
syntax also includes the principles and process to apply the framework into a concrete
land use typology. With that, I argue that beyond the theoretical definition, the practical
context, such as data availability or planning schema will influence the feasibility of a
land use typology.
While we are now in the age of big data, the quantitative change and qualitative change
of data are changing the way that we observe, understand and interpret the world.
While big data provide a changing context for our syntax of land use typology, the
followed case study illustrated one of the possibilities where big data meets the typology
of land use. Use social media data and big data techniques, the case study followed the
syntax of land use typology and developed a corresponding land use typology, which was
applied to two example cities of Boston, U.S and Shenzhen, China.
The result shows big data based land use typology as an innovative description of urban
land use in its focus on human activity, its multiple dimensional land use description, its
flexibility and its speed of data collection.
It should be clarified that based on the syntax, the land use typologies are descriptive
frameworks of land use. The syntax focuses on building a narrative of the current
condition for land use. However, with the syntax, the planners might be able to build
valid land use typologies, based on which they could have a better understanding of the
land use condition. It will help us in the act of planning and administration.
With further consideration, the syntax of land use typology includes a couple of concepts
that are still not familiar to planners and urban researchers, such as the theory about
probability, urban computation, and big data related urban study. To apply these concepts,
we planners should cooperate with computer scientists, engineers and other professionals
to break the barriers.
104
6.1
Value
This thesis seeks not to build a land use typology, but highlights the process of building
land use typology. Rather than fixing the exact classes of land use, the syntax of land use
systematically defines the basic components and possible combinations, therefore reserves
the many possibilities of land use typology.
This thesis implicates land use typology with the newest trend of big data. It also examines
the big data based land use typology as an alternative to understand urban land use.
While big data provide a trigger for the syntax to develop new land use typologies, the
syntax reserves enough flexibility and spaces to go beyond the existing tools and
techniques. The new land use typology could always be developed, such as adding new
elements, new possible packages or new data sources to suit with the fast changing need
of study.
6.2
Limitations of the Study
This thesis is a first-step exploration to develop a syntax for land use typology. For sure it
has limitations in terms of theory, application and practice. In my opinion, three of them
are the most important:
" The syntax only proposes the elements of land use typology, but does not mention
how to order and combine them. It does not put deeper thought on how to
organize the elements exactly when putting them into the land use typology
system.
* The syntax proposes various possible interpretations of the elements and three
generic applications. Yet they are not all tested in the case study, because of my
limited ability. For example, the element of probability is only defined, but not
conducted in both of the two example cities; so is the case for land use estimation.
" The interpretation of the syntax is limited by my scope of knowledge. The
examples given in this thesis, such as land use typologies for energy use, city
105
planning, land management and transportation, are from the fields that are
familiar to me. For other related fields, I was not able to propose the relevant
interpretation of the syntax.
6.3
Future Researches
I would like to propose three aspects of future researches that is meaningful in my
opinion, but not developed in this thesis.
Interpretation of the elements and new schools of packages. The elements and
packages are interpreted according to my understanding of planning. This does not mean
the syntax could only be applied in these fields. We may find it useful for other schools
of research.
Related tools and systems. For example, a database and interactive platform for land use
typology development and illustration. The system could be built based on a database that
includes both the traditional land use data and relevant information from the big data
source. It might allow the users to define the land use elements based on the data source,
develop a corresponded land use typology system with it, and conduct real-time query of
land use information. With the platform, planners will be able to observe according to
their interest and anchor the "hot" area or land parcels that need planning adaption. If
possible, the platform could able be open to the public, allowing them to develop their
own interpretation of land use.
New types of land use. The syntax allows us to combine the elements and develop new
land use typologies. The developed land use typology sometimes might be not valid
because the defined land use types do not exist in the city. But it is possible that these
land use types will come into being in the future. For example, in the case of Boston, we
did not find land use parcels that are in low mix use, but with high intensity. Then we
could ask what it means for a parcel with low level of mix use and high level of intensity,
or whether this type will be valid in the future condition.
106
107
BIBLIOGRAPHY
Agarwal, C., Green, G. M., Grove, J. M., Evans, T. P., & Schweik, C. M. (2002). A
review and assessment of land-use change models: dynamics ofspace, time, and
human choice.
Appleyard, D., Gerson, M. S., Lintell, M., & more, & 0. (1982). Livable Streets.
Berkeley, Calif., etc.: University of California Press.
Bartholomew, H. (1932). Urban land uses. Cambridge, [Mass.] Harvard University
Press, 1932. Retrieved from
http://libproxy.mit.edu/login?url=http://search.ebscohost.com/login.aspx?direct=tr
ue&db=cat00916a&AN=mit.000020782&site=eds-live
Batty, M. (2007). Cities and complexity: understandingcities with cellular automata,
agent-basedmodels, andfractals.Cambridge, Mass.; London: MIT, 2007.
Batty, M. (2008). The Size, Scale, and Shape of Cities. Science, 319(5864), 769-771.
doi:10.1 126/science.1151419
Brown, D. G., Pijanowski, B. C., & Duh, J. D. (2000). Modeling the relationships
between land use and land cover on private lands in the Upper Midwest, USA.
Journalof EnvironmentalManagement, 59(4), 247-263.
doi: 10.1006/jema.2000.0369
Cheng, Z., Caverlee, J., Lee, K., & Sui, D. Z. (n.d.). Exploring millions of footprints in
location sharing services. In In ICWSM2011.
Conway, T. M. (2005). Current and future patterns of land-use change in the coastal zone
of New Jersey. Environment andPlanningB: Planningand Design, 32(6),
877 - 893. doi: 10.1068/b31170
Ginsberg, J., Mohebbi, M. H., Patel, R. S., Brammer, L., Smolinski, M. S., & Brilliant, L.
(2008). Detecting influenza epidemics using search engine query data. Nature,
457(7232), 1012-1014. doi:10.1038/nature07634
Gonzilez, M. C., Hidalgo, C. A., & Barabasi, A.-L. (2008). Understanding individual
human mobility patterns. Nature, 453(7196), 779-782. doi: 10. 1038/nature06958
108
Guttenberg, A. (2002). Multidimensional Land Use Classification and How it Evolved:
Reflections on a Methodological Innovation in Urban Planning. Journalof
PlanningHistory, 1(4), 311-324. doi: 10. 1177/1538513202238308
Guttenberg, A. Z. (1993). The Language of Planning:Essays on the Origins and Ends of
American PlanningThought. University of Illinois Press.
Hartshorne, R. (1939). The nature of geography; a criticalsurvey of current thought in
the light of the past. Lancaster, Pa., The Association, 1939.
Hu, S., & Wang, L. (2013). Automated urban land-use classification with remote sensing.
InternationalJournalof Remote Sensing, 34(3), 790-803.
doi:10.1080/01431161.2012.714510
Iacono, M., & Levinson, D. M. (2011). PredictingLand Use Change How Much Does
TransportationMatter? (SSRN Scholarly Paper No. ID 1735548). Rochester,
NY: Social Science Research Network. Retrieved from
http://papers.ssm.com/abstract=1735548
Jantz, C. A., Goetz, S. J., & Shelley, M. K. (2004). Using the SLEUTH urban growth
model to simulate the impacts of future policy scenarios on urban land use in the
Baltimore - Washington metropolitan area. Environment and PlanningB:
Planningand Design, 31(2), 251 - 271. doi: 10.1068/b2983
Kaiser, E. J., Godschalk, D. R., & Chapin, F. S. (1995). Urban land use planning.
Urbana: University of Illinois Press, c1995.
Kottamasu, R. (2007). Placelogging:mobile spatialannotationand its potentialuse to
urbanplannersanddesigners (Thesis). Massachusetts Institute of Technology.
Retrieved from http://dspace.mit.edu/handle/1721.1/39847
Lam, N. S.-N., & Quattrochi, D. A. (1992). On the Issues of Scale, Resolution, and
Fractal Analysis in the Mapping Sciences*. The ProfessionalGeographer,44(1),
88-98. doi: 10.111 1/j.0033-0124.1992.00088.x
Levinson, D., & Chen, W. (2005). Paving New Ground: A Markov Chain Model of the
Change in Transportation Networks and Land Use. Retrieved from
http://trid.trb.org/view.aspx?id=781603
109
Long, Y., & Shen, Z. (2013). Disaggregating heterogeneous agent attributes and location.
Computers, Environment and Urban Systems, 42, 14-25.
doi: 10.1016/j.compenvurbsys.2013.09.002
Mayer-Sch6nberger, V., & Cukier, K. (2013). Big Data: A Revolution that Will
Transform how We Live, Work, and Think. Houghton Mifflin Harcourt.
Mitchell, R. B., & Rapkin, C. (1954). Urban traffic.
Newman, M. (2006). The Structure and Dynamics of Networks. Princeton University
Press.
Noulas, A., Scellato, S., Lambiotte, R., Pontil, M., & Mascolo, C. (2012). A tale of many
cities: universal patterns in human urban mobility. Plos One, 7(5), e37027e37027. doi: 10.1371/journal.pone.0037027
Paul Mitchell Hess, A. V. M. (2001). Measuring Land Use Patterns for Transportation
Research. TransportationResearch Record, 1780(1), 17-24. doi: 10.3141/1780-03
Pei, T., Sobolevsky, S., Ratti, C., Shaw, S.-L., & Zhou, C. (2013). A New Insight into
Land Use Classification Based on Aggregated Mobile Phone Data.
arXiv:1310.6129[cs]. Retrieved from http://arxiv.org/abs/1310.6129
Pred, A. (1977). The Choreography of Existence: Comments on Hagerstrand's TimeGeography and Its Usefulness. Economic Geography, 53(2), 207-221.
doi: 10.2307/142726
Rise of the digital information age. (2011, February 11). The Washington Post. Retrieved
from http://www.washingtonpost.com/wpdyn/content/graphic/2011/02/1 1/GR2011021100614.html
Schirra, S. M. (Steven M. (2013). Playingforimpact: the design of civic gamesfor
community engagement and socialaction (Thesis). Massachusetts Institute of
Technology. Retrieved from http://dspace.mit.edu/handle/1721.1/81134
Sparks, R. M. (1958). The Case for a Uniform Land Use Classification. Journalof the
American Institute of Planners,24(3).
Sun, H. (2013). The hidden activism: media practicesand the media opportunity in
Chinesepolitics ofresistance (Thesis). Massachusetts Institute of Technology.
Retrieved from http://dspace.mit.edu/handle/1721.1/81135
110
Toole, J. L., Ulm, M., Bauer, D., & Gonzalez, M. C. (2012). Inferring land use from
mobile phone activity. arXiv:1207.1115[physics, Stat]. Retrieved from
http://arxiv.org/abs/1207.1115
Verburg, P. H., Eck, J. R. R. van, Nijs, T. C. M. de, Dijst, M. J., & Schot, P. (2004).
Determinants of land-use change patterns in the Netherlands. Environment and
PlanningB: Planningand Design, 31(1), 125 - 150. doi:10.1068/b307
Von der Dunk, A., Gret-Regamey, A., Dalang, T., & Hersperger, A. M. (2011). Defining
a typology of peri-urban land-use conflicts - A case study from Switzerland.
Landscape and Urban Planning,101(2), 149-156.
doi: 10.10 16/j.landurbplan.2011.02.007
Wang, J., Zhao, M., & Li, X. (n.d.). Discussion on Ideas Concerning Formation Of New
Planning Standards of Development Land In China. City PlanningReview, 36(4),
54-60.
Wilson, A. (2010). Entropy in Urban and Regional Modelling: Retrospect and Prospect.
A V *P
f FX
& FN
: H I1
F
doi:10. 11 lI/j.1538-4632.2010.00799.x
111
. GeographicalAnalysis, 42(4), 364-394.
APPENDIX
A.
1100
American Planning Association Land-Based Classification Standards
Household activities
Includes those activities normally associated with single-family, multifamily, town homes,
manufactured homes, etc.
1200 Transient living
Activities associated with hotels, motels, tourist homes, bed and breakfast, etc. Note that the
distinction between various residential activities is independent of the definition of a family.
1300 Institutional living
Residential living activity associated with dormitories, group homes, barracks, retirement homes,
etc. These activities may occur in any number of structural types (single-family homes, multi-family
homes, manufactured homes, etc.), but the activity characteristics of such living is not the same as
the other subcategories under residential activities. Also note that the distinction between various
residential activities is independent of the definition of a family.
2100
Shopping
Primarily for all retail shops and stores. If the shop sells both goods and services, or if it is
not clear which of the two more detailed categories to assign, then use this one. Increasingly,
distinguishing between a store (that sells goods) and shop (that sells service) will become
difficult and for many planning-related applications even irrelevant. Even economic
applications that employed such distinctions are reconsidering because of the difficulty in
distinguishing between goods and services. However, for those planning applications that
require this distinction, or for existing land-use data sets that already employ such
distinctions, apply the subcategories. Otherwise, for routine land-use data classification,
apply the Shopping category only.
2110
Goods-oriented shopping
Activities in stores that trade retail goods. The distinction is in the physical attributes of
activities associated with goods (buying, selling, repairing, etc.) and not the type of
goods.
112
2120
Service-oriented shopping
Those shops that primarily sell services on site. The distinction is in the physical
attributes of activities associated with services, such as hairdressing. Business services,
such as accounting, legal services, advertising, etc., belong in the office category.
2200
Restaurant-type activity
Eating, dining, and such activities associated with restaurants and other establishments that
serve food, drink, and related products to be consumed on or off premises.
2210
Restaurant-type activity with drive-through
Eating, dining, and such activities associated with restaurants and other establishments
that serve food, drink, and related products that may have seating but has drive-through
facilities. Such activities, although commonly associated with fast-food restaurants, may
also occur at restaurants and food establishments that do not serve fast food.
2300
Office activities
Typical office uses should be categorized here including those that are primarily office-use in
character. Use this category as a catch-all designation for all office-type uses.
2310
Office activities with high turnover of people
Especially those that have counters for customer service, or waiting areas for customers
or visitors. Use this category to indicate an activity characterized by a steady stream of
people when such activity is part of normal operations of the office use.
2320
Office activities with high turnover of automobiles
Typically associated with drive-through windows at banks, department of motor
vehicles, and other businesses. Traditionally, these activities were associated with banks,
post offices, and financial institutions, but they may also occur at other kinds of
establishments.
3j
3100
'Industria. manufactring and waste-retated activities
Plant, factory, or heavy goods storage or handling activities
All industrial activities. Use this as a catch-all category for anything not specified in
subcategories below.
113
3110
Primarily plant or factory-type activities
Assembly plants, manufacturing facilities, industrial machinery, etc.
3120
Primarily goods storage or handling activities
Characterized by loading and unloading goods at warehouses, large storage structures,
movement of goods, shipping, and trucking. Includes self-storage activities.
3200
Solid waste management activities
Includes storing, collecting, dumping, waste processing, and other related operations.
3210
Solid waste collection and storage
Solid waste activities at source or intermediate locations, such as recycling centers. Use
this category for large sites that have their own recycling areas where solid waste is
separated or pretreated. Solid waste includes demolition waste, street sweepings, sewage
sludge, industrial solids and sludges, agricultural manure, and crop wastes.
The term garbage refers to food waste portion of solid waste and refuse or trash refer to
mixed solid wastes.
This category also includes activities associated with recycling (or refuse reclamation)
and other related operations with landfilling.
3220
Landfilling or dumping
Activities that typically occur at landfills and resource recovery facilities. Also useful to
mark those areas not necessarily identified as landfills, but used as dumps. The term
sanitary landfill is sometimes used to differentiate public landfills from others.
3230
Waste processing or recycling
Activities normally associated with incinerators, recycling facilities, resource recovery
facilities, etc.
3300
Construction activities (grading, digging, etc.)
During the construction stage of a development, especially if it is a large-scale one and is a
multiyear project, the characteristics of the use is quite different from what it may eventually
114
become. When local plans need to track such activities, use this category. Once completed,
the activity code should reflect its actual use.
4100
School or library activities
Mainly those associated with educational, instructional, or teaching activities.
Administrative functions, especially those where school board or administrative offices are
located, should be assigned office categories.
Likewise, sports, school-bus parking, or maintenance activities should be assigned
appropriate categories. But if the data being classified is generalizing over large areas, then
use this category.
4110
Classroom-type activities
Those that occur in school buildings, lecture rooms, etc. This category may include
other related activities only if the data is being generalized and the predominant
activities are classroom-type instructions.
4120
Training or instructional activities outside classrooms
Driving, flying, or other instructional activities that occur outside a typical school
building.
4130
Other instructional activities including those that occur in libraries
Include all other instructional activities here.
4200
Emergency response or public-safety-related activities
Broad category to group all fire, police, rescue, EMS, and other public safety activities. Use
this category for joint or co-located facilities if the application needs a single activity code.
4210
Fire and rescue-related activities
The classic example is a fire station with fire trucks in standard bays with associated
training, resting, office, and equipment storing activities on the site. Use this category
for sites that do not necessarily look like a fire station, but serve the same purpose (e.g.,
on-site fire and rescue stations for large-scale developments).
4220
Police, security, and protection-related activities
115
Policing and police-related activities that typically occur in a police station. It also
includes community policing centers located in neighborhoods, which may occupy
store-front locations.
4230
Emergency or disaster-response-related activities
Many look like a typical office building but are distinct in the operations in them. Often
they have the 911 emergency center, disaster coordination facilities, and essential
communication facilities for disaster recovery and response. Note that this category is
not for coding schools and other community facilities used in disaster recovery
operations.
4300
Activities associated with utilities (water, sewer, power, etc.)
Group all utilities: water, sewer, power, gas, etc.
4310
Water-supply-related activities
Category for water supply-related, including irrigation-related activities. Use this
category for any activity associated with water supply.
4311
Water storing, pumping, or piping
Activities primarily associated with linear features, such as pipelines, water
channels, etc., located in easements and point features, such as air vents, pumping
stations, piping junctions, etc., that may or may not be located in easements.
4312
Water purification and filtration activities
Associated with large-scale plants, many of which appear industrial in character.
This category should also include all the related activities associated with a water
purification and filtration facility, such as water storage, water pumping, etc.
4313
Irrigation water storage and distribution activities
This category includes activities associated with urban and rural water distribution
systems. Although not as common as the water purification plants, these activities
are commonly associated with wells and reservoirs for water supply.
4314
Flood control, dams, and other large irrigation activities
116
Associated with dams, reservoirs, and other large-scale storage and distribution of
water. Primarily industrial in character, many such sites also host other activities,
such as sightseeing, power generation, leisure activities, environmental
monitoring, etc.
4320
Sewer-related control, monitor, or distribution activities
This activity is characterized by sewer-related activities, such as pumping, piping,
storing, treating, filtering, etc., whether urban or rural, private or public. Use this
category for any activity associated with sewers.
4321
Sewage storing, pumping, or piping
Activities primarily associated with linear features, such as pipelines, channels,
etc., located in easements and point features, such as air vents, pumping stations,
piping junctions, etc., that may or may not be in
4322
Sewer treatment and processing
Associated with sewer treatment plants, many of which appear industrial in
character. This category also includes related activities associated with a sewer
treatment and processing facility, such as storage, pumping, etc.
4330
Power generation, control, monitor, or distribution activities
This activity is characterized by electrical power generation, control facilities,
distribution centers, etc. Use this category for any activity associated with power supply
and distribution.
4331
Power transmission lines or control activities
Activities primarily associated with linear features, such as transmission lines,
conduits, etc., located in easements and point features, such as air vents, pumping
stations, piping junctions, etc., that may or may not be in
4332
Power generation, storage, or processing activities
Power generation, storage, or processing activities primarily associated with
switching centers, transformer locations, and other power-related facilities that
serve as storage or transit points in the distribution system.
4340
Telecommunications-related control, monitor, or distribution activities
117
Activities associated with telecommunications encompass communication tower
facilities, antennae locations, repeater stations, and distribution centers.
4350
Natural gas or fuels-related control, monitor, or distribution activities
Activities associated with natural gas encompass production facilities, distribution
lines, and control and monitor stations.
4400
Mass storage, inactive
Activities associated with large storage areas for water, fuels, waste, and other products
where such storage is not associated with utilities. These facilities may be associated with a
private or public establishment to serve functions not associated with utilities.
4410
Water storage
Not related to utilities, but may be related to an industrial or commercial enterprise.
This may include tanks, tank farms, open storage, etc., above or below ground.
4420
Storage of natural gas, fuels, etc.
Not related to utilities, but may be related to an industrial or commercial enterprise.
This may include tanks, tank farms, open storage, etc., above or below ground.
4430
Storage of chemical, nuclear, or other materials
Not related to utilities, but may be related to an industrial or commercial enterprise.
This may include tanks, tank farms, open storage, etc., above or below ground.
4500
Health care, medical, or treatment activities
Activities in this category encompass those associated with clinics, hospitals, and other
facilities that treat, house, or care for patients.
4600
Interment, cremation, or grave digging activities
This category encompasses activities associated with cemeteries, cremation facilities,
funeral homes, and the like.
4700
Military base activities
118
Military bases are typically complex collection of activities that include a wide range of
activities associated with military training, living and recreational facilities for military
personnel, storage and maintenance facilities, and other related facilities.
4710
Ordnance storage
Activities primarily associated with storing and moving of military ordnance.
4720
Range and test activities
These activities encompass large areas for range and test activities of arms,
ammunitions, war games, and related military activities. Although such activities are
part of a military base, identifying this special category is useful for planning around
bases for land-use compatibility.
Tlee
5100
Pedestrian movement
Use this category for classifying pedestrian-only roads and open mall areas in
road rights-of-way. Although comprehensive plans may not depend on such distinctions,
many site plans and urban designs use them for circulation components of their plans.
5200
Vehicular movement
This is a catch-all category for all forms of automobile movement on roads, parking areas,
drive-through facilities, etc. Use the subcategories to further distinguish them.
5210
Vehicular parking, storage, etc.
Activities associated with parking or storing of automobiles.
5220
Drive-in, drive through, stop-n-go, etc.
Activities associated with serving customers in their automobiles from a fixed location,
such as a drive-through window. Assign this code to those uses that have drive-through
window facilities. This also includes activities associated with car washes and such
where the customers drive through specialized facilities.
5400
Trains or other rail movement
119
Includes activities associated with movement of rails and other vehicles on railroads. It
includes activities associated with rail maintenance, storage, and rights-of-way for railroads.
5410
Rail maintenance, storage, or related activities
Use this category for identifying rail maintenance and storage activities, which are
industrial in character, from rail movement and railroad rights-of-way. This category
also includes railroad switching activities.
5500
Sailing, boating, and other port, marine and water-based activities
This category includes activities associated with water and marine based travel, movement,
and their related activities. Use the subcategories to distinguish areas of marine movement
from marine storage activities.
5510
Boat mooring, docking, or servicing
Use this subcategory for activities associated with docks and marinas where boats and
ships are anchored, moored, or serviced.
5520
Port, ship-building, and related activities
These activities include a complex collection of shipping, storing, repairing and other
similar activities that are industrial in nature. Passenger terminals are not included in this
category.
5600
Aircraft takeoff, landing, taxiing, and parking
These activities encompass all aspects of air travel and transportation that occur at ground
facilities, such as airports, hangars, and similar facilities. Passenger terminals are not
included in this category.
5700
Spacecraft launching and related activities
These activities include space vehicle control, storage, movement, and viewing areas.
Although they appear similar to air transportation facilities, spacecraft related activities entail
several other activities.
6ass assembly
6100
of people
Passenger assembly
This category is for activities primarily associated with bus, train, and airport terminals.
120
6200
Spectator sports assembly
Spectator sports assembly may occur in stadiums, open grounds, or other venues
occasionally used for such purposes. Identifying such activities may be required for public
safety related applications.
6300
Movies, concerts, or entertainment shows
Besides performance viewing, this category also includes related activities associated with
such performances: food and souvenir vending, purchasing tickets, and related activities.
This category also includes mass assembly at theaters and planetariums.
6400
Gatherings at fairs and exhibitions
Mass assembly of people at fairs and exhibitions includes activities associated with food and
souvenir vending, purchasing tickets, and related activities. This category also includes
activities associated with entertainment shows, park rides, etc., at fairs.
6500
Mass training, drills, etc.
Includes activities in parade grounds and drill fields associated with institutions.
6600
Social, cultural, or religious assembly
Use this category for mass assembly of people for social (eg., city hall), cultural (eg.,
parades), or religious (eg. churches) purposes. It also includes large outdoor ceremonies for
religious, cultural, or other purposes. Although such activities may occur infrequently and
may not involve any functional or structural characteristics (for example a spontaneous
gathering that occurs on an annual basis on a hilltop), identifying where mass assembling of
people occurs is essential for many planning applications. Use this category to capture such
use information. Often this may mean assigning a mass assembly category to areas that
already have other activity categories assigned. Apply this category when other more
specific mass assembly categories are inappropriate.
6700
Gatherings at galleries, museums, aquariums, zoological parks, etc.
Public assembly gatherings at galleries, museums, aquariums, zoological parks, and similar
exhibition services are characterized by a steady stream of people as opposed to mass
congregation of viewers at movie theaters and such. Although the distinction may not be
121
significant, certain public assembly activities require this information separate from other
kinds of gatherings in planning for public safety.
6800
Historical or cultural celebrations, parades, reenactments, etc.
These are usually annual gatherings, parades, and cultural celebrations that may involve
shows, amusement park-like assembly of people, and selling food, drink and souvenirs.
7100
Active leisure sports and related activities
This category refers to an arbitrary second-level coding to accommodate existing data
classified as either active or passive leisure activities. Although the distinction between
active and passive are difficult to separate, use this category only if more precise lower-level
categories are combined in existing data. For new data classification purposes either apply
this category (for top level coding) or identify the precise nature of activities (which are at
the third-level coding).
7110
Running, jogging, bicycling, aerobics, exercising, etc.
Although these activities are normally associated with bike paths, jogging trails,
sidewalks, and such facilities, they also include the kinds that happen on athletic tracks
and playgrounds. Exercising and aerobic activities include those that take place in health
clubs and gymnasiums besides outdoor facilities.
7120
Equestrian sporting activities
This category is for all equestrian-related leisure activities including riding, mounting,
horsemanship, and equestrian games, such as polo, hurdles, dressage training and show
jumping. The related categories include those incidental to maintaining stables, feeding,
caring, and housing horses.
7130
Hockey, ice skating, etc.
This is a broad category to include activities normally associated with ice rinks and
skating on ice. Hockey and other sports on ice are also included in this category.
7140
Skiing, snowboarding, etc.
This is a broad category that includes leisure sport activities on snow: skiing, luge,
bobsled, toboggan.
122
7150
Automobile and motorbike racing
This is a broad category to include the myriad forms of vehicular sports including
automobile racing, dirt racing, motorcycle racing, and other cross-country type events.
7160
Golf
Includes other leisure activities, such as pall-mall, tipcart, croquet, golf, curling, and pall
one besides golf.
7180
Tennis
Because of its unique site development characteristic, traditionally lawn tennis (as
opposed to table tennis) has been classified distinct from other sporting activities. It also
includes related sports, such as racquet ball.
7190
Track and field, team sports (baseball, basketball, etc.), or other sports
This includes activities associated with playing baseball, basketball, and other related
games.
7200
Passive leisure activity
This category refers to an arbitrary second-level coding to accommodate existing data
classified as either active or passive leisure activities. Although the distinction between
active and passive are difficult to separate, use this category only if more precise lower-level
categories are combined in existing data. For new data classification purposes either apply
this category (for top level coding) or identify the precise nature of activities (which are at
the third-level coding).
7210
Camping
Camping is a broad category that includes parts of activities associated with of shelter,
recreation, and other related activities, such as hunting, fishing, sailing, etc. The
designation applies to only those camping areas and camp grounds where camps are
allowed.
7220
Gambling
Casinos normally host gambling, wagering, and those establishments that serve the
gaming aspects of leisure activities. However, many other types of establishments also
123
provide slot machines, and other gambling and gaming facilities (shopping centers in
Las Vegas, for instance).
7230
Hunting
Hunting activities include live and also clay pigeon and skeet shooting.
7240
Promenading and other activities in parks
This is a catch-all category for all other areas of parks and recreational areas that do not
qualify under any of the other more specific categories.
7250
Shooting
7260
Trapping
7300
Flying or air-related sports
7400
Water sports and related leisure activities
7410
Boating, sailing, etc.
7420
Canoeing, kayaking, etc.
7430
Swimming, diving, etc.
Includes activities associated with lifeguard services and other related activities.
7440
Fishing, angling, etc.
7450
Scuba diving, snorkeling, etc.
7460
Water-skiing
80l
8100
s
sIN
Farming, tilling, plowing, harvesting, or related activities
Agricultural activities, such as farming, plowing, tilling, cropping, seeding, cultivating, and
harvesting for the production of food and fiber products. Also includes sod production,
nurseries, orchards, and Christmas tree plantations. Excludes forest logging and timberharvesting operations.
124
8200
Livestock related activities
Activities associated with feeding and raising of livestock in pens and confined structures.
8300
Pasturing, grazing, etc.
Activities normally associated with feeding and grazing in open ranges.
8400
Logging
Activities normally associated with forestry.
8500
Quarrying or stone cutting
Includes activities normally associated with borrow pits.
8600
Mining including surface and subsurface strip mining
Includes crushing, screening, washing, and flotation activities. Beneficiating is another
common term used to describe such activities.
8700
Drilling, dredging, etc.
Includes activities normally associated with on and off-shore drilling for oil and natural gas
operations, dredging for beach control, expanding waterways, and cleaning of canals or
channels.
9100
Not applicable to this dimension
Use this code as a permanent code for those records that will never be classified in this
dimension. It is normal for land-use databases to have records that may never be classified
and be left blank instead. But LBCS recommends that all records have a code because some
computer applications may not be able handle blank entries (null values in database
terminology).
9200
Unclassifiable activity
Use this category as a temporary placeholder for activities that cannot be grouped anywhere
until the classification scheme is updated. Check the LBCS web site to see how others have
dealt with such unique activities before revising the classification scheme.
125
9300
Subsurface activity
Use this category for activities that occur below the surface that are of no interest to the
applications that will use this data set and assigning one of the unknown categories may be
inappropriate.
9900
To be determined
Use this code as a placeholder until an appropriate code can be assigned. It is normal for
land-use databases to have records that may never be classified and left blank instead. But
LBCS recommends that all records have a code because some computer applications may
not be able handle blank entries (null values in database terminology). This code could also
be used as the default value for data-entry work. The subcategories serve the same purpose
for other coding levels.
B.
China National Standard of Land Classification Gb/T 21010-2007
*tF
012
J
..
...
.-_.-.__- - ----
..
.....
.
01
126
02
N
i
021
fh"A,
M)K*soflnff,00CM* 4,AM, iM%-Aft M.
NA
022
03
023
J4
031
T
032
j
*A
fLrH"
tem
-K&
#- M Ve
41
k) A
R--n+I
VA 4 M I
fA" J-1 9
033
.....
5 >.26 5 * ...
04
*......
* *t
041
)
PHW
59-<0%-IPM'
A
r,
M
)4:
-)UTMVANMj
042
0
43
11I
$w1I
127
----------
05
------------
itt
~I1
:tt~
9T1
052
06
053
j
054
4
--------
----------
T
-@~~061
TA1
t~~z~
J~k
-----------------
062
063
-1L:
128
07
Ninai...............
JM
071
ft
072
08
a
Ji
fN
f
082A
il
---4--------------i--if m
0814J )[1
Tt-P)Meisi
Avl
085
129
086
&a
A
087
fiWAAMAJi
jlAIEjntk01AfHf-A5
Apil MIT
088
Rx
41-if T RS l ,t
.t
A
f4NA
09
JA
092
093
_
-----------4IRM-T-LIM,
130
m
095
10
ki
iA4
ifiuk
101
t9z
PRtA-i
fflo- tIRitI)mIm
& m , MA t
tfL
~1TJ
102
mni
5Mu
will
0t'5
*MunilmI. wai
nit
14
1025~
mtaitn'
106
1647
om.
*J
l1ti& zomii
04PtiiiVS4J
4
f>)
X
'F1
4T
tA~itlisW
AK
'FR4
vil)msmee
------------
#4
10
4
)
131
107
it
107--------
---TQ
------
1
----
jA
11
7~
Th~
k~J1~
1121,+p
A
lt)
M(J flEx 11 wdff A f
112
-------
116
kj-L
115
4vm.1fJf~
-4
a
----
132
119-
)11J1~
12
it
121
i
122
~ilk
ilL
Ak
LII
123
At
124
kk
,
R
Th
V'
~
Om :bJ2.Om flJX
tP
125
126
W}kA.
Tf 4
127
±1k 7O%MJA±
----- ---
AA#RTII±1k
133
----
------
XKANW
201iffsin
------2
-
ONa, Wt, TI _L_%
nWWenki
-----
e
#A,
fttt T19", TAN
t4;ft;(tiW
203-Ar
C.
ta,
*A
AL
TW-i-
A)1), *t
ff
'_tA
China Code for Classification of Urban Land Use and Planning Standards of
Development Land GB 50137-2011
X @ k4!
W2.0.1 J+
a.
ng lis) xw
cod es
Chinese
English
H
development land
non-development land
* 3.2.2 AT
M ifP
I UL
~~IJ fA~Oj~
I-i
4rt, T1-44111
[VJ~ ~
III
iQfrT 1i~,
[HI
Aiitid
H- 12
AMi, it
H
13
H-14
A'f , fAl
;, f;jl)i-.Olal , 10, 3-±--i QM ILI
NandIINenm
MI ,Ir iA
IZ91
±iM
fTkE Id
134
f YEft IL M M,
tA-UnnN
t t
MKjff1
A~ Lz N
1H21
tk-Y ffILL'
1-123~
1-125
d
IJ
WW4FE
ay wtJ'J
uil i~pd, 1- i
lI4+AE[PII. iJ4JFiE
ThjAL-K VT}Yd! I~' A-1I~2IfR
XIU
mTi~
alfKfit,
h
4-% E
F141J1
i Iffi V 'W
H42
~~ Iii tij~~ ift, -4 MI~
i? q, fILI
Eli
MALd~1 Cll, AJ141
A T.
,"E , iil*A L, -Le
E-12
101Tjn3T
itil
135
IIR}l
~' {"]' F 10 JJ M.3
it
fl -)Wy
M,(J w
q
L I--
1r
1Y; Vfw
n Z
tininr
N ht4A
E31
f4l I
ItzBLL~A~J~4
E32
A 2.0.2
MA
codes
Chinese
English
p7
residential
A
administration and public services
tfi A
B
commercial and business facilities
M
industrial
logistics and warehouse
Rit
S
street and transportation
U
municipal utilities
G
green space
3.3.2
±
f ft{ Q
Rk~
RI
R11
R12
R2
TtfA
t-t9, A,
1JAm
E
T4 M14T0 t MP 4:*R ,
9hf74
waneest,
136
fletm
R21
~ fl
R22K~jk~ffl4
R3 &i)
A34
A
It ExI l
fl f
~
a it wA35
lIj-4J~
a
.
Akikkl W
iM
EJL
tAfRhl
f*
137
I
A4
k-------
A5
-4
A5
f
-
MA
A53
A59
A6
-14Y
A7
a
mPTit
B
Bl
k akffl
-
-- - - -------- --- -
B1
B12
13
B2
B2
11- Af9R N ff
|
t
I1
AMR
B22
AIM,
B329
R
&R*Affle
Nfl
O
I
I
&A&
LI4~
4<
B3
rxj MI,j l j*f*q'i kl IMIILJ
B31
WfJdL
J,
T6
5%M TIi
138
Qfit l t
lKfiit)
B4
mi2 6A
A
B49
t~f, fl- fth
J4 fi)]j]{
B9
m
T~~~T
ml
T R it] if-
-
J
M2
V-~ -61 Ak qjt. 3hJF ,11ffM J1 tP- {I 1Jil N It~ llf
tft
4,f~
4QITJX
~
f'
1tT ifflA1 xkti f2T ff 1, 4,
,
M3 _ ~kj 4 11 RI' kiJ~Thf{U2~ YLJEVYJ
RR JL
W2
kI-ft
ff
T4~l1
M, nF~ IuIA M1 I E
i'P NTll kA'L rJij Alit
T t (14i4
ff
mn
PHT)
4J41J
I
JIP
W3>J
SI
S2
JiiftiIP
ht
S4
S41
-~i AL LQ lfI b b K{i M ft1
~ff
hl5i
NM
119
APt
YIP
N?
~)it] [1
4
L It
- tA t1Ai
A>l - ti Lb M Jtl i bk
139
f1AKP.
~~J'~
9-4-Tl.-*fi
i
l ,I
J1
flfJP. 0,J YIP.) Mf
A 4ViiL
~I 1-4) A:
GIL JP.V
fYYi1iQ IP.
jRIY j fl tP.
7A tL IIftYM,
LHLb
U115
I]i
Th T
Ni1r±t'Jj m~)~
i
U11
LJ22
T, 11iQ rp-ff iht
U }tI vP)Lfl vi t f1I fdi
U32E ff It
f*#IfHKJ hVI-M --Hth
A~h m
fp)M
h-Alot,
jl)iff~ t
ilirYM,
-
AWR
J
~ [~FJ11~.ITh~Jfhuu; ~Yijyf~IFth
Liit3
f~i
~~
NX Nit]
Jhk~~~~P
j'A
TfAL
rI14
140
f4ljH}n-tF,
I Af I
I-Tt~~
t44I
J
MWIfI
PJi1R
141
M MI RYl
nl Fi /
tiJit 4 f It