Banner: Yay, transit!

Using GTFS Data in ArcGIS Network Analyst

Add GTFS to a Network Dataset - Sample Analysis

Chicago library hotspot analysis

This Story Map shows a proof-of-concept analysis identifying hotspots in Chicago where citizens have a high level of need for better walking/transit access to public libraries.

To produce this map, I first created a network dataset for Chicago using Add GTFS to a Network Dataset. Then, I used this network dataset in the OD Cost Matrix tool in ArcGIS Network Analyst to calculate the travel time between each census block centroid and the nearest public library.

However, because the available transit service changes rapidly throughly the day, the travel time can be drastically different. In order to factor in this heavy time dependence, I ran my OD Cost Matrix calculation with a start time of each minute between 12:00PM and 5:00PM (the hours when all the libraries are open). I wrote a python script to automate this calculation.

To summarize all the OD Cost Matrix results, I calculated the percentage of times during those hours when each census block had access to a library by transit or walking in under 15 minutes of travel time. I factored in demographic data to calculate a level of need score for each census block. Blocks with a higher level of need score would benefit the most from better access to public libraries. I used the Optimized Hot Spot Analysis tool to identify blocks with statistically significant levels of need.

Cincinnati supermarket accessibility

The map below shows a measure of transit accessibility to grocery stores in Cincinnati, Ohio.

To produce this map, I created a network dataset for Cincinnati using Add GTFS to a Network Dataset. I used the Network Analyst OD Cost Matrix tool to calculate the transit/walking travel time between each census block centroid and the closest grocery stores. Because transit travel time varies throughout the day, I calculated this OD Cost Matrix for every minute of the day. Using these results, I calculated the percentage of times throughout the day that each census block had access to at least one grocery store and mapped it using a color scale.

To learn more about this analysis I did with Steve Farber and Michael Widener, view the poster we presented at the 2013 GIS in Transit conference and the Applied Geography paper we published.

Sample Analysis: Cincinnati grocery store transit accessibility

Analyses by other people

I appreciate your questions, comments, and suggestions.
mmorang@esri.com