Click the screenshot below for a scrollable, interactive map of residential tax parcels in New Rochelle, a suburb just north of New York City. The colors indicate the assessed value per square foot of the land, from purple (low) to green (high), split into 10 deciles.
One inference which we can almost certainly draw from this data: property taxes collected from New Rochelle’s somewhat rundown neighborhoods near downtown — the mostly-green area on the southwest side of the map — are subsidizing the more expansive properties to the north side of the city. (Though, to be fair, I don’t know anything about what proportion of city tax dollars come from property taxes versus other sources.)
I was inspired by the visual appearance of a similar map of home values in Tyler, TX (though I chose to refocus on value per square foot, the appropriate unit for a constricted urban area). I used MapBox for hosting and TileMill (a MapBox product) for creating the map tiles.
The underlying map is the Open Street Maps Bright tiles, which I figured out how to use thanks to this MapBox tutorial. The GIS data for the New Rochelle overlay comes from the generous folks of Westchester County. The tax data comes from a huge, unwieldy pdf file produced by the city of New Rochelle as part of their assessment process. I used pdftotext to put it in a marginally workable form and then stripped the data I needed in Python, using the dbfpy library to merge the tax records with Westchester’s GIS shapefiles.