- Spatial Data Modeling
- What is a model?
- representation ~ simplification ~ abstraction
- something that represents the 'real thing'
- model car
- flow chart
- formula
- diagram
- map ...
- reduced detail & only selected features/information ...
- 4 Levels of Modeling (abstraction)
- 1 Real World - in all its complex glory
- 2 Conceptual Model - your understanding of it
- 3 Logical Model - possible computer realization (software independent)
- 4 Physical Model - actual digital structure (software implementation)
- Conceptual Model
- 2 "Views"
- Discrete
- features that occupy space
- entity types
- points (well)
- lines (creek)
- polygons (watershed)
- geographic space
- filled with discrete features (objects)
- features have location & attributes
- features have relations with other objects
- nearby? - well near creek?
- connected? which creeks are connected?
- within? creeks within watershed?
-
- features to map?
- entities used?
- Continuous
- surfaces
- the value varies through space
- examples of surfaces
- elevation
- slope
- soil depth
- rainfall
- sound
- geographic space
- change in location ... change in value
- no hard boundaries
- Issues
- How should Nanaimo be depicted on a map? or the Fraser River?
- discrete vs. field
- forest
- trees or stands (Belize example)
- if stands
- ...what about height/ volume uphill?
- thus a forest can be modeled with
- points
- polygons
- surface
- could you model it with lines?
- wetland
- degree of wetness/ habitat value
- what about wetland within the forest
-
- Vector
- spatial entities - coord.
- points
- distinct objects (wells) vs ...
- point values (spot hts)
- lines
- linear object (creek) vs ...
- isoline (contours)
- polygons
- defined by boundary
- area within is considered "homogeneous"
- topology
- relations are determined and stored in database
- connectivity (networks)
- adjacency (neighbours)
- containment (within)
- allows for error checking of line wrok
- overshoot
- undershoot
- open polygon
- attribute data
- in tables
- JOINED to spatial data via unique feature ID
- e.g. forest stands
- Raster
- grid of squares
- resolution - size of pixels (edge in real world units)
- each pixel (cell) is independent
- point, linear & polygon features still possible
- excellent for cont. variables
- attribute data ... in the map ... in the pixel ... like a speadsheet
- Raster vs. Vector
- advantages & disadvantages
- vector
- advantages
- accurate boundaries
- visually pleasing
- great for discrete features
- database table allows many fields of data
- topology possible - relational analysis
- network and flow analysis
- smaller file size
- disadvantages
- data structure is complex
- cannot "analyze within a polygon"
- plus opposire of raster advantages
- raster
- advantages
- excellent for continuous data
- satellite imagery analysis possible
- terrain modelling straightforward
- overlays are easier
- disadvantages
- large cell size ... blocky
- small cell size ... large files
- often need a map layer for each single attribute
- plus opposite of vector advantages
- Vector = Discrete ... Raster = Field?
- how can vector model a field?
- how can raster model objects?
- Physical Model
- Vector
- basics
- features defined by coordinates
- lines/ poly's ... connect the dots
- coordinates are stored in data tables
- attributes are stored in data tables
- unique ID provide link btwn feature table and attribute table
- spaghetti model (3.15a)
- each feature stored independently of others
- all shared boundaries stored twice
- no relationships recognized
-
- relations btwn map features are defined
- relations: adjacency, containment, connectivity
- allows logical data checks
- Raster
- basics - as above
- file format - just a long list of values
- requires a 'header' or 'world file' of meta data
- # rows/col's
- georeference system
- extents (usu. top left coord.)
- units & resolution
- min/max values
- header/ world file converts file to a 'geo-spreadsheet'
- common image formats
- JPEG
- TIFF
- GIF
- BMP
- GIS rasters
- GRID (ESRI)
- GeoTIFF
