Lecture 10: An Overview of Spatial Database Development: the project development approach Introduction This course is designed around three blocks of material: 1) database design; 2) GIS database development; 3) GIS analysis and display. This lecture marks the beginning of the second block of material which will concentrate on data acquisition, processing, and integration. As an introduction to database development, today we will discuss the GIS Project Development Approach (PDA). PDA is a formalized process aimed at developing GIS projects. The three components of the PDA are Determining Project Objectives Database Design Database Development Today’s lecture will look at each of these components but concentrate on database development, providing an overview of the next 5 weeks. Determining Project Objectives Before developing a GIS application, the objective, or problem that the GIS will address must be determined. Although selecting an objective seems like a no-brainer, GIS projects often fail because this step is either skipped or mismanaged. This is especially true if there are multiple opinions, or if the project is highly unstructured. There are a number of methods available to help in determining project objectives. We will look at two examples: Rich Picture, and Root Definitions. The Rich Picture method uses a schematic graphic of the issues necessary to determine a project objective. The schematics include defined graphic conventions to represent areas of conflict (crossed swords), outside viewpoints (eyeballs), both official and unofficial positions (a variety of text balloons), etc. These schematics allow the presentation of the main components of a problem, as well the interactions between the main players. Rich pictures can be drawn by individuals or by groups. Rich pictures tend to represent a single view but at the same time allow for multiple viewpoints. In the example to the left, the rich picture presents a variety of problems, interactions, and views of a proposal for a real estate GIS. RNR/GEOG 417-517 Gary L. Christopherson 1 Root definitions involve multiple statements submitted, negotiated, and edited until there is a single, agreed upon statement. Sticking with the real estate example presented above, the root definition of a real estate agent might be expressed in the following statement: The GIS is a system to help maximize home sales; the root definition of the home buyer as: The GIS is a system to help identify and rank possible homes; and the negotiated, compromise root definition: The GIS is a system that identifies homes for sale that meet the requirements of the buyers. In addition to these, there are a number of other techniques, structures, and software tools that can help identify the objective of a project. Database Design The second component of the GIS Project Development Approach is database design. Having covered this topic ad nauseam during the first 9 lectures, we will pause only briefly here to discuss design issues specific to the PDA. Successful database design will create a structure that allows the project objective/s to be met. Given that the combinations of objectives and available data sets are unique, database design will nearly always be unique as well. For example, database design for projects dealing with Forestry and Administrative Boundaries are very different in structure and components. Each is designed to answer specific questions, using specific data and specific analytic techniques. This is not to say that these designs are useful only within the projects they were created for. They can be used as templates for future products, edited and changed to become new designs suited to new, unique projects. In addition to the project objective/s, essential issues to consider when beginning the design process are project constraints and nature of the final product. Understanding what the final product will look like (analog/digital, journal article, web site, etc.) will have a significant impact on project design. Constraints on the project are also important. For example, data constraints may mean that your data is not adequate to meet objectives; you may not have a large enough budget to support the design; the people you’ve hired may not be trained adequately to carryout the development of the design; etc. RNR/GEOG 417-517 Gary L. Christopherson 2 Database Development Database development constitutes the majority of time and money spent on most GIS projects. Because of its importance, the next several lectures will cover a variety of database development topics, including: Data inventory Data acquisition/automation Data integration Data transformation Data conversion As we do a quick survey of each of these today, you will notice that there is a lot of overlap between these topics. For instance, data transformation will appear in discussions of inventory and integration, as will data conversion. This indicates that database development is not a strictly linear process, but goes forward and backwards in order to develop a database that matches the objectives of a project. Data Inventory The data used in a spatial database can be both varied and extensive; knowing what you have and don’t have is an important first step in database development. Data for a project can range from physical to analog to digital; from maps to text to remotely sensed imagery; from GIS to CAD to GPS; from 1:100000 to 1:500 to 1” = 1 mile; from DMS to Conic to UTM; from good to bad to indifferent. The importance of a data inventory cannot be over estimated. Without an inventory essential data could be overlooked, duplicated, or used inappropriately. Data inventories should include information about: Source Format Scale Resolution Projection Automation type Level of integration Maintenance RNR/GEOG 417-517 Gary L. Christopherson 3 Data Acquisition/Automation Data automation is the process of converting non-digital data to digital data. Automation uses a variety of technologies, including digitizing tablets, scanners, GPS, and total stations (below) to capture data directly or to convert data from analog to digital formats. Traditionally automating data was the most common form of data acquisition and dominated spatial database development. As readily available spatial datasets become more widely available it has become less common but remains an important aspect of project development. For students learning about GIS it is particularly valuable as a pedagogic device for understanding vector data structures in a GIS. Beginning next week in lab, you will be automating data in maps to GIS layers in a geodatabase. Digitizing Tablet Large Format Scanner Using a GPS to Collect Data Using a Total Station to Collect Data Integration Perhaps the main reason that GIS exist is their ability to combine different datasets for analysis and display. Before this can happen, integrating the data layers is necessary. This integration takes a variety of forms the most important of which are geometry, attributes, projection, and format. Each is important for analysis; in fact spatial analysis will fail if one of these lacks integration. The easiest way to describe this is by taking a quick look at geometric, or vertical integration. Vertical integration is registering features from different RNR/GEOG 417-517 Gary L. Christopherson 4 themes, but with common boundaries, to a common location. In the graphic above, state boundaries do not match county boundaries. Using these data sets in an analysis of shoreline development would create problems because each layer would produce a different shoreline and thereby a different analytical outcome. In the same way attribute integration is about making sure that attributes in different data sets, representing the same information, are defined in the same way and that the values populating these attributes come from the same domain. Data Transformation This aspect of database development concerns projective transformation of spatial data. That is, the transformation of spherical data to a flat surface. This process will distort spatial data in terms of shape, area distance, or direction. As seen in the image to the left, different projections produce different distortions. Each has its own strengths and weaknesses with some being better suited to particular analyses than others. Spatial analysis should not and often cannot be carried out on data in different projections. Because of this, it is important that spatial data be normalized in terms of projective transformation. Data Conversion The final topic in our survey of database development is data conversion. This is the process whereby data in one file format is converted to another file format, usually one more suited to use in a particular GIS. Although many GIS software applications will read data in other formats, this often limits the capabilities of the software for data management or analysis. For example, ArcGIS will read AutoCAD files but the structure of these files limits them in key areas such as topological awareness, data management, and analysis. For these reasons, it is advisable to convert all data to a format that is fully functional in your GIS application software. Summary This lecture used the components of the PDA to provide an introduction to the GIS Project Development Approach. It concentrated on database development as an overview of the material that will be covered in the coming weeks. RNR/GEOG 417-517 Gary L. Christopherson 5