Introduction to Geoprocessing in GIS Flashcards
What are the 6 steps in the conceptual approach to problem solving in GIS?
1) State problem
2) Break problem down
3) Explore input datasets
4) Perform analysis
5) Verify Results (repeat step 4 if necessary)
6) Implement
Model Builder
Proprietary to ESRI
- Can also use R or Python
- Used to make suitability analysis more efficient
Stating the problem
Don’t go further or analyze
ex. Need new highschool
Breaking the problem down
- Objectives (away from other schools, has a student population, bus stops, flat topo, available land plots)
- Data needed (schools, transits, demographics, zoning, topography contours/DEM)
Exploring input datasets
- Relationships
- Whats in the data i.e. metadata (when was the census?)
- Look at data
- Look at distributions (queries & tables)
- Reclassifying data to a similar scale to help aggregate
Performing Analysis
- GIS Tools
- Ex. buffer fro other schools, euclidean distance (raster), bus stops euclidian distance
- Weight which is the most important (close to one but away from another)
Verify Analysis
May have to return to Performing the analysis until it is right
- Output should be range of suitability fro lowest to highest
Why do we use model building and geoprocessing?
- Automation (for many files)
- Data manipulation
- Work flow (Multiple-step procedures)
- Data transformations (project, clip, buffer, etc.)
- Helps to sketch the model first on paper
- Power is in the ability to link together many of these data transformations (Model!)
What does it mean when the features in a model are not coloured in?
Not coloured means the links aren’t connecting properly and that there are errors and breaks
Example of a basic raster suitability analysis
- Select data layers
- Analysis (buffers, slope, aspects, ranks, etc.)
- Reclassify attributes to same scale
- Combine layers (weighted overlay)
- Constraints (0,1)
- Final suitability map
Why do we reclassify attributes to the same relative scale (ex. 0 - 100)?
- To normalize the data scale
- To help aggregate data
- Convert vector to raster for a weighted overlay
- we don’t want to be comparing 13m to 27m
Why do we do a Weighted overlay?
Weight inputs based on which are more ‘important’
- Rank each layer based on its suitability for final layer
AHP
Analytical Hierarchy Process
- Justifies approach to choices but doesn’t tell that the weights chosen are perfect
- Tells how “well/consistent” the pair-wise comparisons are
- Shows that workflow is good, not exactly that the weights are good
- Compares criteria pair by pair
- Comparison and judgement of alternatives in a pair-wise manner
What is the scale for AHP? - Relative language (slightly, more, etc.)
1/9 = Completely dominated by 1/7 = Far less important 1/5 = Less important 1/3 = Slightly less important 1 = Equal importance 3 = Slightly more important 5 = More important 7 = Far more important 9 = Completely dominated by
What is the pinch point?
Weighting
- Different weights will give different outputs
