A few days ago Taz pointed me to the fact that Census 2010 had been releasing a lot more data in the past few months. I was naturally curious, so I decided to check out the website. Mind you, I’m someone rather familiar with the older website, and have downloaded raw data sets and crunched them with R. So I was cautiously optimistic. Very cautiously. Apparently the big news is that the American Factfinder now has a “web 2.0” version which will be releasing 2010 count data. Unfortunately, the implementation of AJAX makes the site very, very, slow (and beware of non-supported browsers!). And, there isn’t that much 2010 information.
To review, there are decennial censuses which are straightforward counts, and, since the aughts there have been American Community Survey results which are based on a sample, and so have a margin of error (this makes county-level data less useful because for small subpopulations within a county the margin of error can be very large). I’m looking forward to 2010 results because they are going to be much more robust than the ACS data which makes the news periodically for small subgroups. You can see where I’m going with this, since the core readership of this weblog comes from a group which is itself subdivided by nationality, ethnicity, class, and religion.
Though the initial intent was to find 2010 results for Indian Americans which I could compare across the older censuses, I stumbled onto some interesting language data. The pie chart is based on 2006-2008 ACS data. I’ve put a
csv file of the data online. For the pie chart I removed some languages with less than 1,000 claimed speakers. The original data can be found here. The sample was limited to those aged 5 and over. If these data are correct ~2.25 million Americans spoke indigenous South Asian languages at home in the second half of the aughts. Continue reading →