News for hackers. Stuff that matters.
Re:The Racial Dot Map: One Dot Per Person for the Entire U.S. by araes on Thursday September 11, 2014 @01:12PM (#8303747)
So many great data analysis exercises that could be run with such a set. Effectively, you could create scalar variables representing lots of different population statistics on a country, state, city, ect level.

Density: http://i.imgur.com/bfVdrbK.png

Racial Density / Clustering: http://i.imgur.com/pQDJLtX.png

Terrain / Feature Clustering: http://i.imgur.com/AH9f2KW.png

City Homogeneity: http://i.imgur.com/Lo3X6KI.png

City Racial Mixing: http://i.imgur.com/VBzLPI6.png or http://i.imgur.com/ldIgYaw.png

City Cluster Homogeneity: http://i.imgur.com/2TLabMj.png vs http://i.imgur.com/LlsAMx4.png

Cluster to nearby dot gravity (how much do dots spread from a nearby loci): http://i.imgur.com/LlsAMx4.png vs http://i.imgur.com/qNkMCcu.png

And that doesn't even cover stuff that's harder to show like gradients, edge analysis (who's likely driving between two towns?), or migration patterns if multiple census years are available.

There are no comments on this comment yet.