Posts

Showing posts with the label demographics

Demographic Analysis with… Google Street View? And #DeepLearning!?!

Image
The car you own is a good predictor of your income, education, race, and voting pattern. Oh, not   you   in particular, of course not! But analyze enough of you — say, across 200 cities in the US — and the predictions turn out to be eerily accurate. Gebru et al. do exactly this in this excellent paper  —  https://goo.gl/HbnK34  — using cars extracted from Google Street View images, 50 Million of them across 3068 zips and 39,286 voting precincts. The results themselves aren’t as important (°) as the way they derive it. Cars need to be classified by make, model,   and trim level   (can   you   tell the difference between a 2007 and 2008 Honda Accord? Based on a tweaked tail-light?) something that ends up involving 2,657 buckets —  #DeepLearning   really excels at this. Also, given that this is all just from Street View images, you could, plausibly, do the same with other factors such as house spacing, type, and, heck, the shrubbery   (And ...

Race & Sex on American TV - a #Visualization

Image
Amptoons puts Alyssa Rosenberg's count of the race and sex of characters on prime-time TV up against the same for the USA as a whole.  The conclusions? Asian & Latinos get short shrift  Law Enforcement is all over the place Check it out...

The U.S. Melting Pot - How many of you are there?

Image
From Bloomberg, an interactive map allowing you to not just see how many of you there are, but even more importantly, how you compare against others - e.g., Indians vs Polish, Mexican vs Puerto Rican, etc.  Its absolutely brilliant - go check it out !

Interracial Marriages - a Pew Study

Image
The Pew Research Center has a study out on interracial marriages , and the numbers are fascinating. First, the high point (literally), 15.1% of all new marriages in 2010 were 'mixed race', up from 6.7% in 1980. Or, to put it differently, of all married people , 8.4% are mixed race, as compared to 3.2% in 1980. And, as you can see from the lines to the left, its pretty much a straight line and shows no realy signs of flattening out. Now for the wacky number.  If you break the couples down, the highest earning couples are White/Asian by a pretty significant amount (almost $8K/yr more than the next highest group - Asian/Asian).  Weirdly enough, this pairing is not really all that common, with only the White/Black pairing being less common. Of course, the best part of the study is that the taboo has most certainly faded.  From the study As of 1987, two decades after the Supreme Court ruling, just 48% of the public said it was “all right for blacks and whites t...