About this site

Subscribe to the RSS Feed

2010 Census: Leveraging what’s already been done

Part of what I do at work is make maps. To me, the best maps are those that show you, at a glance, something interesting about what’s going on around you, and, if you’re curious, allow you to explore what’s going on in the rest of the world. I’ve been following what the Tribapps team has been doing with the 2010 census data as it has been coming out and after going to a recent Hacks/Hackers Meetup where Joe Germuska presented some of the APIs that they’ve been working on and I gotta say, I was inspired. Below are some preliminary steps and observations.

Custom boundaries, normalized data

During the day, I work for the Bahá’ís of the United States. The administration of the affairs of the Bahá’ís across the country are organized based on geographic boundaries which we call, originally enough, Bahá’í Localities. The boundaries of these geographies sometimes match up with normal things like counties, townships or municipal boundaries but are more often that not combinations or subdivisions of those. The other complexity is that they combine together to form other types of administrative subdivisions which have nothing to do with any other existing shapes defined by people who have large warehouses full of geographers busily drawing lines on maps. The upshot of all this is that if you want to try to project something like census data onto these areas, it can get kinda tricky. Luckily, someone had the foresight to assign a FIPS code to each of these areas so at least approximations can be made.

Also luckily, someone (namely these guys) has already done a bunch of the dirty work when it comes to normalizing the data provided by the census as well as putting a JavaScript API on it. For me, however, since I’m going to need to do a lot of custom combinations of that data, and to be quite honest I don’t really need all of it, I’m thinking I’m probably going to need to pull down the data that I need at put it into the Django models that house the rest of our geospatial data. That way I can preprocess a lot of it and have it ready when it’s needed.

Preliminary steps

So far all I’ve done is setup the Django project powering census.ire.org and poked around with the API a bit. There are a ton of scripts within the project which do all kinds of wonderful things to get going with the mountain of data within the census. Since I’m mainly interested in grabbing this data and comparing it to what I’ve already got established with my larger mapping project, the first step is to actually make the models. I’ll try to keep this blog up to date with any pitfalls I encounter or brilliant discoveries I make. So, uh, if you’ve read this far, stay tuned.

blog comments powered by Disqus