Automating Analyses with ModelBuilder

1
Automating Analyses with ModelBuilder
2
Overview

• Why Use ModelBuilder?
• ModelBuilder Basics
• Common ModelBuilder Problems
• Advanced ModelBuilder

3
Why Use ModelBuilder?

• An automation tool
• But comes with some startup overhead
• Most useful in two circumstances
• Documents models their parameterization
• Allows experimentation with model parameters
  particularly for weighting and rating
• Common Types of Models
• ETL Extract, transform and convert raw data
• Suitability Building attractiveness maps

4
ModelBuilder Basics

• Basic idea is that of a dependency diagram
• User specifies inputs, processing and outputs
• If inputs change, system repeats intermediate
  operations as needed
• Diagram has three kinds of elements
• Inputs
• Geoprocessing Operations
• Outputs
• Output from one operation can be used as input to
  an other, allowing chaining

5
ModelBuilder Setup

• Rather obscure to startimplemented as a custom
  toolbox tool
• Open toolbox panel
• Create empty toolbox
• Right mouse on Toolboxes, select New Toolbox
• Create empty model
• Right mouse on new Toolbox, select New Model
• Then populate model by drag and drop
• Of data layers from map table of contents
• Of geoprocessing operations from the toolbox
• Finally, wire data and processing boxes together

6
Example Simple MB Model

• Goal
• Create a factor map expressing simple proximity
  to residential landuse where output is classed
  from 1..9
• Method
• Create new model
• Select residential landuse from San Miguel
  Parcels database
• Add Euclidean distance geoprocessing operation
• Connect landuse (input) to distance (process),
  specifying new grid (output)
• Run
• Add Reclass Operator
• Connect output grid of distance operator to input
  of reclass, specifying new output grid
• Run again

7
Review of Simple Model

• Benefits
• Multiple logical steps encapsulated in a single
  step
• Model Logic Recorded in Diagram
• Model Parameters Recorded
• Problems / Caveats
• Default is not to show results
• Model as Created is 100 specific to particular
  data paths/locations on disk
• Model Saving Bizarre..
• Default operation names make no sense to end
  users
• Spatial Analyst Toolbar Options do not inherit

8
Showing Results

• Simple, but not Obvious
• Right Mouse on Output -gt Add to Display
• If at first you dont succeed, try toggling again

9
Saving / Finding Models

• By default, models saved in My Toolboxes folder
• Main menu Tools-gtOptions-gtMy Toolboxes
• Default is C\Documents and Settings\(Username)
  \Application Data\ESRI\ArcToolbox\My Toolboxes
• Easiest to find in ArcCatalog/My Toolboxes
• Can Add Toolbox stored on disk

10
Making Models Generalizable

• Running Models
• Can Double Click on Models in Toolbox Panel
• By default, not too useful, because no user
  control of outputs
• Generalizing Models
• By default, models only use exact data originally
  specified
• To make a model into a true tool need to
  specify which inputs / outputs are variable
  parameters
• Right mouse on input or output
• Select Parameter (toggle)
• After Parameters are set, double clicking brings
  up user dialog

11
Making MB Diagrams Legible

• All elements can be renamed from right mouse
  menu
• Rename layers if necessary to clarify
• Explain intent of geoprocessing operations
• i.e. Isolate Residential Landuse instead of
  reclass1
• If Desired, change diagram properties
• Square Circle Square
• Box Background Colors
• If you need better quality, export diagram

12
Environment Variables in MB

• Note
• Spatial Analyst Options settings not inherited
• Must explicitly specify for MB
• Two options
• Can do once for all toolboxes (recommended)
• RM Top Toolbox-gtEnvironment Settings
• General Settings -gt Extent
• Raster Analysis Settings -gt Cell Size
• Can do once for each model

13
Model 2 Weighted Overlay

• Goal
• To Create an Attractiveness Model with ability to
  Weight factors
• Method
• Create separate ModelBuilder models for each
  factor
• Nest models into master MB model
• Combine with weighted overlay

14
Model 2 Implementation

• Factor 1 Proximity to Residential
• Factor 2 Proximity to Ski Slopes
• Created by copying and pasting factor 1 model and
  adjusting inputs and outputs
• Weighting
• Factor 1 2X Factor 2
• Use Spatial Analyst Weighted Overlay tool