Geospatial Data Structures GEOG3222: Geographical Information Science II Dr. Jed Long Department of Geography Western University Lecture Overview Lecture Overview • Review definition of geospatial data • Geospatial data representation • Data models: objects and field • Vector • Raster What are Geospatial Data? What are Spatial Data? • Spatial data are a collection of measurements taken at specific locations • Data that are mappable http://www.ci.fort-collins.co.us/gis/what-is.php Components of Geospatial Data Components of Spatial Data 1) Location – describes where a “thing” is 2) Attribute – provides information about http://www.ci.fort-collins.co.us/gis/what-is.php the “thing” Components of Geospatial Data Components of Spatial Data Examples Tree Location: A single tree Attributes: • Tree height • Tree species • Tree health Components of Geospatial Data Components of Spatial Data Examples Healthcare • Location: hospital location • Attribute: • Hours • Specialties • Patients http://gis.broward.org/gisdata/hospitals.htm Representing Geospatial Data Representing Spatial Data • Need some way storing real world phenomena in a digital format. • First consideration …. How do we represent spatial phenomena? Representing Geospatial Data Representing Spatial Data Data representations impact how we: • Conceptualize • Analyze • Interpret Representation & Conceptualization Representation & Conceptualization What are the fundamental properties of geospatial data? Two fundamental properties Properties of “things” in the real world The spatial relationships between “things” Objects and Fields Data Models • Objects are things that are mapped using discrete boundaries • Fields are continuously varying across geographical space OINT Field OBJECT Representing Spatial Data ENTITY ENTITY FIELD FIELD Data Models (conceptual) Data Representations (practical) POINT LINE LINE VECTOR AREA AREA RASTER RASTER TIN TIN Data Models Objects and Fields Philosophical argument that dates back to ancient Greece about the nature of reality: empty container full of distinct objects or continuously varying field of phenomena Object Models Object models • Object model = entity model • Space is conceptualized as a collection of selfcontained objects and relationships between objects • Phenomena under study are self-contained objects • Objects described by attributes Field Models Field models • Considers that phenomena have spatially continuous attributes • A value is possible at an infinite number of point locations on a surface • Note – pencil method to determine field or object OINT Field OBJECT Representing Spatial Data ENTITY ENTITY FIELD FIELD Data Models (conceptual) Data Representations (practical) POINT LINE LINE VECTOR AREA AREA RASTER RASTER TIN TIN Object Representations ENTITY Points Lines Polygons POINT LINE AREA Spatial Points • Pair of coordinates • 2-D • Polar coordinates • Latitude, Longitude • Cartesian coordinates • { X, Y } • Maybe some attributes • { X, Y, Attributes } Soil Samples: Zinc Concentration Spatial Lines Spatial Lines represent linear features, often called polylines. • Examples: • Roads • Rivers • Power-lines Spatial Lines • Polylines = poly (many) + lines • Collections of connected coordinate pairs • First point not same as last • { (x1,y1), (x2,y2), … , (xn, yn) } • Typically, non self-intersecting… Spatial Polygons Spatial Polygons represent areal features. • Examples: • • • • • Continents Lakes Parks States/Counties etc. Cities/Urban area Spatial Polygons • Polygons are collections of points • First point = end point • “close the loop” • Can have holes, etc. • Discrete = “floating” • E.g., Lakes • Lattice = “connected” • E.g., States/counties Field Representations ENTITY FIELD • Rasters • Lattice • Tins (will talk about later) LINE AREA RASTER TIN Raster Space is divided into smaller units Space is tessellated A tesselation is defined as the process to cover a surface through the repeated use of a single shape. Raster A raster can use any reasonable geometric shape, as long as it can be connected in such a way as to create a continuous surface. Grids • Square pixel rasters are the most common • Easy to deal with mathematically • Efficient to store Raster Data Raster data represent continuous spatial variables with a regularly patterned spatial unit (e.g., pixels). • Examples: • Elevation (DEM) • Satellite Imagery • Land cover Vector to Raster Choosing Representation Should the data be represented with object of field? Often the “thing” being studied could be conceptualized as either E.g., Forest polygons • Continuous forest coverage • Groups of similar tree - stand Representation Weakness: #1 • Multi-temporal spatial data • Change through time Representation Weakness: #2 • Uncertainty • Many natural processes are fuzzy • A forest patch is both naturally easy to think of as an object, a collection of objects (trees) and a continuously varying field