Linked by Thom Holwerda on Mon 7th Apr 2014 19:55 UTC
Talk, Rumors, X Versus Y

I'm using the URL slug headline for this one (check the link).

This map showing the locations of 280 million individual posts on Twitter shows a depressing divide in America: Tweets coming from Manhattan tend to come from iPhones. Tweets coming from Newark, N.J., tend to come from Android phones.

If you live in the New York metro area, you don't need to be told that Manhattan is where the region's rich people live, and the poor live in Newark. Manhattan's median income is $67,000 a year. Newark's is $17,000, according to U.S. Census data.

This fascinates me, as it seems to be a very American thing. In The Netherlands, Android has an 80% market share, and we have far lower poverty rates than the US (that Newark median income is crazy low by Dutch standards). I'm pretty sure the situation is similar for many other West-European nations.

This raises an interesting question: is it 'Android is for poor people' - or is it 'Android is for poor people in America'?

Permalink for comment 586539
To read all comments associated with this story, please click here.
Not this again
by derekmorr on Mon 7th Apr 2014 20:35 UTC
derekmorr
Member since:
2005-09-25

This is old "news" -- it first ran last summer. It's BS. Classic data visualization bug -- the site draws Android data points first, then draws iOS data points on top of them thus hiding many of the Android data. If you bring up two browser windows, side-by-side, and toggle one to only show iOS, and the other Android, you get a very different impression. I did some side-by-side screenshots when this first ran (and Danny Dilger at AppleInsider picked it up) - http://forums.appleinsider.com/t/158143/twitter-heat-map-shows-ipho...

In addition to the visualization problems there also appear to be problems with the underlying dataset: http://cartonerd.blogspot.com/2013/06/3-billion-tweets-on-map.html

In short, Twitter use is skewed towards young, urban males, and very few users geo-tag their tweets. I wonder how and if these biases were controlled for in the analysis? Just plotting lots of data points on a map is not analysis. It's map porn.

This story highlights several problems I have with the popularization of so-called "Big Data." It can lead to lazy analysis. Too many folks seem to think that volume of data makes up for sloppy analysis and questionable sampling. It doesn't.

Reply Score: 16