Learn spatial analysis techniques with scenario-based case studies

The Applied Analysis team has been hard at work developing scenario-based, cross platform exercises to help you learn spatial analysis techniques for solving your real-world problems.  Each case study includes data and step by step workflows for multiple ArcGIS applications, such as ArcGIS Pro, ArcMap, ArcGIS Online, and Insights for ArcGIS.  Most of the case studies discuss key analysis concepts and lessons such as how to deal with missing data, how to explore and take advantage of the spatial structure in your data, how to model supply versus demand, how to transform data, and how to select an appropriate weighting scheme. Work through the case study analyses using the data provided, and then adapt the workflows to your own data.

Exploratory analysis

Exploratory analysis is a technique that helps you explore and learn more about the patterns and distributions in your data.  Exploratory analysis can either stand alone as a complete analysis, or be used as a starting point for another type of analysis, making exploratory techniques some of the most essential and ubiquitous techniques in spatial analysis. Exploratory analysis is used in almost every industry, including health, public works, retail, and natural resources. Here are some case studies that use exploratory analysis techniques:

Suitability analysis

Suitability analysis is one of the most commonly used GIS analysis techniques, and for good reason.  This type of analysis has broad application for a diverse set of industries, including economic development, urban and regional planning, retail, manufacturing, environmental management, and public safety. While the problems may be diverse, the analysis follows the same basic steps: determine a set of suitableBobcat criteria, apply the criteria to your data, and derive results that fit that criteria. Here are some examples of case studies that use suitability analysis:

Cost distance analysis

Cost distance analysis is a technique used to determine how a traveler should move through a landscape based on a set of resistance criteria.  The result of  cost distance analysis is a least cost path – a path of least resistance between a start and end location – or a network of paths that optimally connect a set of desired locations.  Cost distance analysis is very popular in sectors related to wildlife because it allows planners to design the best configuration of wildlife corridors to allow the species to move between habitat patches.  Cost distance analysis can also be used for other applications, like creating routes for new roads, trails, electrical lines, or pipes, as well as identifying the best movement paths for fire fighting or military operations.  Here is an example of a case study that uses cost distance analysis:

Analyzing spatial clustering

When we look out at the world around us, we see clustering: disease outbreaks around tainted water sources affect aid distribution, shoppers Glass bottlescluster in space and time creating economic opportunities, and there are crime and traffic accident hot spots affecting public safety.  Analysis of clusters and clustering is important for a large variety of industries including elections and redistricting, urban and regional planning, insurance, retail, real estate, banking, telecommunications, water resources, defense, education, health and human services, transportation, and public safety.  The case studies below explore statistically significant hot and cold spot areas in space and time, or rank locations based on their similarities and differences:

  • Analyzing traffic accidents in space and time – A spatial and temporal analysis of crash data in Brevard County, Florida based on a road network and accounting for variations in commuting patterns.
  • Analyzing violent crime – A spatial and temporal analysis of Chicago’s violent crime in relation to liquor establishments and unemployment.
  • Locating a new retirement community – Uses supply versus demand modeling and the factors associated with the most successful communities to narrow potential locations for a new community.

Analyzing correlations

Is average income a good predictor of luxury TV purchases?  Do student-to-teacher ratios help explain test score results?  We might believe that agricultural areas with the best growing conditions will produce the highest yields, but is that always the case everywhere?  If not, why not?  Understanding relationships between and among data variables is powerful because it is a first step toward explanation and, ultimately, encouraging positive changes.  These methods can be applied to any application area where you want to better understand the factors promoting spatial outcomes.  The case studies below use correlation and regression analysis:

  • Modeling literacy – Explores key explanatory variables contributing to literacy rates in Africa and uses cluster analysis to suggest targeted country level remediation programs.
  • Mapping the geography of online lending – Using regression analysis, finds surprising discrepancies in the expected relationship between average loan grades and average interest rates.

Creating surfaces

Statistical or physical, 2D or 3D, surfaces are an important component of many GIS analyses.  Surfaces such as elevation models, temperature maps, or density heat maps are typically modeled as continuous raster surfaces, created from points or contours usingSoil interpolation or geostatistical methods.  Surfaces are used in analyses associated with a number of different industries including forestry, weather and climate, environmental management, sustainable development, and mining.  Here are some examples of case studies that use different geoprocessing tools to create surfaces:

Network analysis

Routing is probably the most commonly used GIS technique by non-GIS users; everyday people are using maps to find the best route from their home to the store, or from their hotel to the museum.  Network analysis takes it a step further to solve complex routing or location problems with complicated sets of variables.  Don’t worry though, we’ll walk you through the workflows step-by-step so you’ll barely even notice.  Some typical network analysis applications include routing for mail and package delivery, emergency vehicle routing, and location-allocation to site store warehouses or outlets.  Here are some examples of case studies that use network analysis:

Get started!

What are you waiting for?  Look over these case studies to learn some new approaches for analyzing your data. Keep in mind that all of these case studies reflect common workflows that can be generalized to other data and other applications.

Don’t see a case study with your most common workflows?  Contact us!  We want to help you with your analyses and to write up your story as a case study that will help others!

Are you an instructor?  Feel free to use these case studies for classroom exercises.

Photo credits

This entry was posted in Analysis & Geoprocessing, ArcGIS Online, ArcGIS Pro, ArcMap, Spatial Statistics and tagged , , , , , , , . Bookmark the permalink.

Leave a Reply

3 Comments

  1. Brian Hovis says:

    Thank you for the compilation of different exercises for spatial analysis techniques. I have done some projects using these techniques, but your blog inspires me to do more.

  2. Brian Hovis says:

    Vicki, I wonder if you can steer me in the right direction. I want to look for multiple precincts (voting districts) that consistently vote for the same party over an eight year period. The trouble is, the precincts change a little over the years so I can’t compare like to like. I would like to find blocks of precincts that vote in certain ranges. I am starting by plotting the data as you suggested in an earlier post. I will then look at tools that find clusters.

    If you have any insight, I would appreciate your thoughts.

    • Vicki Lynn Cove says:

      Hi Brian,

      There are a few options for how you could solve this problem, but the main thing you have to do is convert all of your data into a single geometry and then stick to that geometry. Depending on the details of your analysis, it may make sense to use the most recent precinct boundaries, or you may want to pick a different type of boundary, like counties or census tracts. Either way, you will have to apportion your data to the boundaries you’ve chosen.

      Here are a few ways you can apportion your data:
      1) Use aereal interpolation from the Geostatistical analyst extension. You can check out the documentation for aereal interpolation as well as this video to get you started.
      2) Use Spatial Overlay or the Data tab from the Business Analyst extension.
      3) Use Summarize Within in ArcGIS Pro or ArcGIS Online. There are also other summary tools that are available in ArcMap.

      Once you have apportioned your data into your boundaries, you can use the statistical method of your choice for determining voting patterns. You may want to do a hot spot analysis, or if you have ArcGIS Pro, then the space time pattern mining tools are very cool, especially emerging hot spots and visualizing the space time cube in 3D. If you use the space time pattern mining tools then you will want to use the defined locations version of the space time cube tool. I believe you must have at least 10 time intervals to create a space time cube.

      I hope this information helps. Let me know if you have any other questions.