The shapefiles, which contain the land area (used for density) as well as a lat/long combination, are found on the Census website
To download all of them at once, run:
bash wget -m ftp://ftp2.census.gov/geo/tiger/TIGER2014/TRACT/
And the zip files will be downloaded in a nested folder.
To unzip all of the data files at once, run:
bash find ./ -name \*.zip -exec unzip {} \;
This unzips .dbf files, which are a type of database file that can be converted to .csv using csvkit and the function in2csv with a -f dbf flag. We output the transformation into a file using:
bash find ./ -name \*.dbf -exec in2csv -f dbf {} \; > file.csv
So file.csv has some repeating lines (the headers), but those get ignored when we merge the file with data about population.
The population data came from the Minnesota Population Center, which allows for historical data and geographies to be downloaded, although it does require creating an account. I downloaded this file and saved it as population.csv
0. Data prep details the cleaning and merging of the two data sets.
1. Distance vs. Density plots details the plotting. Future work will attempt to fit a model to the data or allow for user input.