Fun Statistical Data Sets?

November 7, 2010 8:55 PM

I'm looking for data to use in a statistics class. The data set needs to have less than 25 variables and at least 200 individual records. Any fun/unique/silly ideas? I need to right a paper analyzing this data -I'd like it to be interesting. I need access to raw data - it is okay if it has already been analyzed but I need to show that I can analyze it.
posted by turtlefu to grab bag (11 answers total)

HERE IS THAT THING YOU WERE LOOKING FOR.

Welcome to Data.gov. If you're into this sort of thing it might just be the last time we ever see you.
posted by The White Hat at 8:59 PM on November 7, 2010 [3 favorites]

Less than 25 variables, n = 200? Can you cut down from a bigger set? That's really small.
posted by k8t at 9:06 PM on November 7, 2010

If you're at all interested in social science, check out ICPSR. They collect high quality social science data sets, and their data downloading interface is super user friendly.
posted by sashapearl at 9:06 PM on November 7, 2010 [1 favorite]

The Pew Internet & American Life project has some data sets that might be fun.

A bit macabre perhaps, but the US Fatality Analysis Reporting System data is really fun to play with.

If you want to analyze he economic activity in your area the Census's County Business Patterns database is cool.
posted by otio at 9:44 PM on November 7, 2010

The metafilter infodump perhaps?
posted by Iteki at 9:45 PM on November 7, 2010

R is chock-full of sample data sets.
posted by Blazecock Pileon at 1:03 AM on November 8, 2010

Maybe some of the Ig-Nobel prize winners have published their data. Or you could try to unravel the math behind the famous scientific hoaxes.
posted by Psychnic at 3:28 AM on November 8, 2010

Less than 25 variables, n = 200? Can you cut down from a bigger set? That's really small.

Yeah, what I did in this circumstance was cut out a subset of the most recent National Election Study. GSS would work just as well.
posted by ROU_Xenophobe at 5:14 AM on November 8, 2010

Thanks for these. Just to clarify n > 200. And I can shrink a larger data set. So far it looks like data.gov is fitting the bill as I can download directly into SPSS.
posted by turtlefu at 7:05 AM on November 8, 2010

Of course, there's always baseball.
posted by bluejayway at 9:09 AM on November 8, 2010

IPUMS.
posted by oceano at 10:50 AM on November 8, 2010