I looked over the Netflix developer API site briefly I don't think this is public data on rental trends, so Netflix must have collaborated with the article author to show us all something interesting. Also, you have to take into account this si RENTAL data... so a movie may have been extremely popular in the theater for a zip code's population, but very seldom rented.
Here is an example from the article map mash up. This is a picture of the movie, "Vicky Cristina Barcelona" and it's popularity in the Bay Area. The more red a zip code, the higher it rated on the collective Netflix queue's of that zip code's population. If a zip code is white, the movie did not appear in the top 50 movie queues for that zip.
So this is already pretty interesting. Based on what demographic information I know about the area simply by living here, I can easily WAY over-generalize this and say this movie was vastly more popular with Caucasians. Of course the more meaningful analysis is when you can interpret the other direction, saying zip code XYZ is going to be interesting in movie ABC. Movie studios spend tons of money on this kind of data. Even more meaningful is when you can infer some things about one map, use that inference to make politically incorrect over generalizations, then apply those same over generalizations to another set of zip codes.
More interesting are other movies which are telling on other dimensions. I'm not very politically correct anyway, so I'm happy to suggest that this picture of rental trends from the movie, "Milk" tells us interesting things about public opinion of homosexuality:
Especially as compared to say, the same movie and it's respective popularity in Dallas for example:
Or this picture of the Bay Area, which counter-intuitively tells us the movie, "The Proposal" was not very popular with well educated single people:
