The analysis performed on this site includes data from a number of sources. Various city, county, and state voter records, the U.S. Census, the American Community Survey (ACS), and the Citizen Voting Age Population (CVAP) tabulation are all used in order to provide deep insight into municipal voting patterns. To start, voter data were compiled into a consistent format in order to facilitate accuracy. Then, along with Census, ACS, and CVAP data, voter data were uploaded into ArcCatalog for spatial analysis.
To analyze the data we generated a number of ArcGIS models. The details of these models are too lengthy to discuss here, but the first step prior to using them involved geocoding the acquired and prepared voter records. We used the geocoder from ESRI's ArcGIS Business Analyst to create a point on a map for each voter record.
The first models select 2014-vintage Census tracts in the vicinity of our study area (city limits) and append Census data tables to the tract shapefiles. Once complete, the next set of models spatially join voter data to these tracts and aggregate the voter data based on the desired variables. In other words, they select each voter within a given tract, and the end result creates a table with the total number of voters falling into each category (total registered voters, turnout in an election, voters falling into our specified age categories, and so on).
After compiling tract data, the models calculate the aggregate city-wide numbers. The end result of these models provides us with a feature class in a GIS geodatabase containing all of the Census tracts in a city meeting our criteria with all of the required data attached.