Geog 463 GIS Workshop by wuxiangyu


									   Geog 469
 GIS Workshop

Data Analysis - 1

  April 20, 2010
1. What is a typology of GIS operations?
2. How do we use workflow process diagrams?
3. What is geovisualization?
4. Discussions
   1. Typology of GIS operations
• Analysis is the core of GIS that transforms data into

        INPUT          OPERATIONS        OUTPUT

• Data manipulation operations (Tomlinson functions)
  – data availability  data readiness
• Data analysis operations (Tomlinson functions)
  – data ready to put into system  information
              GIS operation context
After Figure 13.6 (Longley et al p. 296)                         Validation,

Various steps in workflow                         Analysis

                                 & manipulation


 Real world                                Many kinds of analysis
    Needs for data manipulation

                                                        Image source:
                                                        Bolstad 2005

• Incompatible cell sizes and boundaries confound multi-layer raster
  operations. These ambiguities are best resolved by resampling prior
  to layer combination
              Data manipulation
• Data creation
   – Digitizing
   – Vectorization (ArcScan)
• Data editing
   – Topology editing (only in Arc/Info coverage and geodatabases)
   – Raster resampling (coregistering different raster images)
• Data conversion
   – Export/import (be aware of data loss due to incompatibility)
• Conversion among coordinate systems
• Coordinate transformation
   – Registration (control points)
   – Transformations
   – Rubber-sheeting
 Examples of data manipulation
• Coordinate             • Raster resampling

               • Shift
        Data analysis typology
•   Queries
•   Measure
•   Transformations
•   Descriptive summaries
•   Optimization
•   Modeling

                            Following Longley et al 2005
             Queries (extract)
• Attribute queries
    Select * from table where [condition]
    Condition: logical expression
    e.g. Median Age > 45

• Spatial queries
    Select * from layer where [condition]
    Condition: based on spatial relationship
    e.g. hospital within King County
    e.g. hospital 10 kilometers or farther from highway
• Measure distance, area and parameter of a
  single feature
  – ArcToolbox or calculate fields
  – Use meaningful measurement unit (by converting
    geographic to state plane)
  – Make the use of ArcScript to measure more complex
    geometric properties on the ESRI website

• Measure slope and aspect of terrain
  – Work with DEM
  – ArcGIS Spatial Analyst
      Transformations (vector)
• Simple spatial analysis that change datasets,
  combining them or comparing them to obtain
  new datasets
• Buffer
• Spatial overlay
  – Mostly translated into point in polygon operations
     • e.g. # parcels within zoning area?
  – But there are an array of operations for polygon in
    polygon overlay with the added complexity (union,
    intersect, identity)
• Other commonly used
  – Clip, dissolve, erase, ….

      Transformations (raster)
• Map algebra for raster
  – Local function (also known as reclassification)
  – Neighborhood function (focal function)
     • filtering
• Spatial interpolation
  – Deterministic (specific function) : IDW
  – Stochastic (probability) : Kriging
  – Use Geostatistical Analyst for those
          Descriptive summaries
• Capture the essence of a dataset in one or two numbers
• Mean and standard deviation of attributes
   – It may not be included in the process diagram, but useful in
     gaining a overview of data and determining the appropriate data
     classification method for thematic mapping
• Descriptive summaries of spatial features
   – Mean center, dispersion (use CrimeStat)
• Measure spatial pattern of vector data
   –   hot spot detection methods: nearest neighbor, K function…
   –   Work with “point” data
   –   ArcToolbox - Spatial Statistics Tools
   –   Try CrimeStat by National Institute of Justice (freeware)
• Problem-solving by optimizing objectives
   – That is, minimize cost or maximize benefit given constraints
• Find the least cost path
   – Finding the best path based on link-node representation where
     cost information is attached
       • Arc/Info workstation – dynamic segmentation, other software such
         as TransCAD, ArcView Network Analyst
   – Finding the gradient path based on grid data
       • ArcGIS Spatial Analyst – least cost path
• Location-allocation problems
   – Location for a new retail store?
       • Arc/Info workstation – location-allocation
     Need for spatial modeling
• In reality, it is impossible to experiment on the
  real world
• Instead, the model provides alternatives to
  experimenting on a digital replica of the world
• Allows us to evaluate the outcomes of different
  policy alternatives (i.e. “what-if scenario”)
• Can be used for dynamic simulation, providing
  decision makers with dramatic visualizations of
  alternative futures
       Spatial modeling defined
• Definition 1: A sequence of different operations; aka.
  cartographic modeling (by Tomlin); process diagram
   – What if you want to change parameters and see what happens
     to outcomes? Want to automate the whole process?
   – Try Model Builder in ArcGIS: it allows you to run the whole
     procedures with one click
• Definition 2: Process model (extracted from previous
  knowledge) is embedded in operations so as to predict
  what would happen given input parameters
   – Range from simple to complex
      • Soil erosion = f (a, b, c, d); housing price = f (a, b, c, d)
      • Hydrological modeling: sequence of logics or rules
   – Test what-if scenarios
• Seek out models developed by experts
• Perform literature review to identify published models
    Example: evacuation plan
• Identify high-risk area which is likely to be
  congested in the event of wildfire given a
  limited amount of time
• Congestion = f (network configuration,
  population density, weather, other)
• What if we build the road with different
  design? What would happen?
• Planners can use the outcome of this
  simulation to design the road
Defining characteristics of modeling
• Involves multiple stages
  – How something will change over time

• Incorporate the previous knowledge
  – Approximation of reality
  – Some assumption on circumstance

• Experiment with scenarios or input parameters
  – Can compare different outcomes from varying
    scenarios or parameters
    Multi-criteria decision modeling
• Many decisions require the evaluation of multi-criteria
• Different stakeholders commonly hold different
  “values” about what is important
• Values can be translated into criteria (factors)
• How should that importance of criteria be measured?
• How can various criteria measurements should be
• Example: problems of hazmat transport
   – Save money, save human life, save environment, save all
     three to what degree? to arrive at agreed alternative?
Conflict resolution in multi-criteria method

• Don’t negotiate alternatives
• Negotiate criteria, for example, what should be
  included as important criteria? and what weight
  should we provide to those criteria?
• We can performance sensitivity analysis on
  alternatives, criteria, and weights to arrive a
  mtually agreeable outcome
• Software: ArcInfo, IDRISI from Clark Labs,
  CommonGIS, GeoChoicePerspectives - all have
         How can you get help?
• Help file
   – Search by commonly used operation names
   – ESRI Library (manual)
• Software website
   – User Support, User Forum
   – Make the use of available scripts
• Previous work
   – Draw upon literature reviews
   – Web of Science for academic works
   – GIS magazines for professional works
• UW-gis-l list
   – To subscribe,
• Instructors
2. Workflow process diagram
    Using ModelBuilder
    Nature of Transformations
In previous diagram
  buffer, buffer and recode are based upon
  spatial data transformation operations …
  - buffer: line to polygon operation
  - buffer: line to polygon operation
  - recode: point to point operation
What is the basis of these transformations?
See…Nyerges and Jankowski 2010 RUGIS
  Spatial data transformation tables
  we look at these in more detail in next lec
Sequencing Transformations
Determining criteria (factors)
              Refining criteria
General      Refined criteria   Refined criteria
criteria     (binary scheme)    (continuous scheme)
Slopes not   Slopes < 30
too steep    degrees

Far enough 300 meter < X
from road to distance to road
provide      < 2000 meter
privacy, but
not isolated
Determining weights
Combining the criteria using overlay
    then recode to a ranking
   3. Geovisualization defined
• Create and use visual representations to
  facilitate thinking, understanding, and
  knowledge construction about human and
  physical environments, at geographic
  scales of measurement

                               From MacEachren
     Geovisualization purpose
• Foster greater user interaction and participation
  in the use of GIS as decision support tool

• User can undertake a balanced appraisal of the
  message of a geographic phenomenon;
  decision-makers can appraise the
  consequences of different representations

• Increasing role of users
   Geovisualization and maps
• Why do we use maps?
• Can we utilize a more “interactive”
  environment within GIS?
• Facilitate thinking, understanding, and
  knowledge construction
• Roles that a display can play
  Map presentation vs. Data exploration
          Roles of a display
 Map presentation vs. Data exploration
                    Map presentation   Data exploration
                    (presentation      (Geovisualization)
Interaction between Low                High
map and user

Public or private   Public             Private

Information sharing Info sharing       Knowledge
or knowledge                           construction
  What are advantages of geovisualization
     compared to map presentation?
• Query
   – You can make a simple geographic inquiry
   – See how message is sensitive to input data elements
• Transformation
   – Data can be manipulated such that objectives can be served
     better; it helps you overcome the limitation of traditional
     cartographic products (e.g. choropleth map) such as
   – Cartogram: areal size adjusted to magnitudes
   – Dasymetric map: spatial overlay
• Immersion
   – Expand the capacity to experience the world through virtual
     environment (3D representation of the world)
   – Data-rich environment (Lidar), Computing power (animation)
   – Facilitate distributed decision-making process
      4. Discussion questions
• Why do we build model given the fact that
  the model is an approximation of reality at
• What would be challenges and issues to
  consider in spatial modeling?
• How is geovisualization different from
  presentation mapping?

To top