The application fills a need for people who enjoy walking to find routes that provide the most satisfying walking experience. We especially designed the application to allow users to weight ten walkability attributes in order to find a route that optimizes their choices.

Applications like Walkscore and Walkshed provide useful information about the conditions of a particular area or neighborhood that facilitate walking trips. The overall walkability of the area is scored on a scale indicated by a number (walkscore) or by a color coded map (walkshed). Walkshed does allow users to adjust preferences related to walkability to suit their needs. However, none of these applications have the ability to offer actual routes that users can follow to get from an origin to a destination. Our application is designed to provide alternative walking routes based on user preferences.

There is an extensive literature examining the factors that affect walkability of an area or neighborhood. Several characteristics of the built environment have been found to influence an individual’s propensity to walk. We have reviewed the essential parts of this literature in our publication: Guhathakurta, S., G. Zhang, M. Pangaluru, and R. Sivakumar (2013). Walk Route: A new methodology to find the optimal walking route in the city of Atlanta. In Chapter 18: Planning Support Systems for Sustainable Urban Development, edited by S. Geertman, Fred Toppen and John Stillwell. Heidelberg: Springer. Given extensive literature on the attributes of walkability and their associated measures, the selected attributes for this application fall under ten categories. These are: 1) residential density; 2) business density; 3) land use diversity; 4) land use accessibility; 5) street connectivity; 6) crime safety; 7) traffic safety; 8) physical barriers; 9) aesthetics; and 10) pedestrian infrastructure.

The data used in the application are collected from various sources, most of which is publicly available. We used the street network data from 2007 Georgia Department of Transportation and the Atlanta Regional Commission street file. Block boundary data were obtained from 2010 U.S. Census Bureau TIGER/Line shapefiles. A measure of intersection density was calculated using the ArcGIS network analysis tool based on the street network data. The variables in the residential density group were calculated using the ArcGIS operator tool based on the 2010 U.S. census block data. Similarly, the variables in the business density, land use mix diversity and land use mix accessibility categories were calculated using the ArcGIS operator tool, zonal analysis tool and Euclidean distance tool, based on the 2012 reference USA business database. Crime data were obtained from 2009-2011 police department report, provided by the Atlanta Police department. Slope was calculated using the ArcGIS 3D analysis tool, based on the 1999 digital elevation model (DEM) data (30 meters pixel resolution) provided by the U.S. Geological Survey (USGS). We used ERDAS 9.3 Indices tool to calculate NDVI (A simple graphical indicator that can be used to measure vegetation coverage. The higher the value of indicator, the higher the vegetation coverage) based on the 2011 Landsat image data provided by the U.S. Geological Survey (USGS). Traffic volume data were obtained from the 2009 average annual daily traffic (AADT) data provided by the Georgia Department of Transportation. Public transit data were collected from the 2010 Georgia regional transportation authority file. Finally, sidewalk data were obtained from the Georgia Department of Transportation.

We will continue to update the data as the organizations and agencies who provided our initial datasets release their subsequent later version. For most of the datasets we are using this could happen every two-to-three years depending upon the type of data and the agency from which it is sourced. Some data can be updated yearly or sooner (e.g., crime statistics) others could take 10 years (census enumerations). We are continuing to find ways to update our information in real time -- as they get collected by various agencies.

Yes, our algorithms and methods are already documented in the paper noted above. We want this application to be a template for future applications developed for other cities.

We are continuing to develop this application in different fronts: a) We are developing methods for user input about the suggestions provided by this application and the actual user experience; b) our immediate objective is to port the application to mobile platforms (Android and iOS); c) we are working on developing better methods to collect, update, and verify the data required for this application; and d) Variants of this application are already being discussed that would cater to other user needs such as for those who are wheelchair bound or have other mobility impairments.

Yes-- in the works!

For everyone who loves the walking experience and would like to find new paths to the same destination!