Do You Belong In Your Zip Code?

The New York Times recently published a web service showing the movie rental habits, by zip code, of 12 major cities over the past year. The data comes from Netflicks movie rentals for 100 movies. I initially visited the site just to see how they presented the information (very well) and it really made me think. This is a great example of how data is only what you make of it. And, that no matter how unbiased you try to be, just collecting and presenting data induces inacuracy or bias. The act of interpreting it also requires generalizations, risking bias and bad judgement.

My first thought was, "Wow, this is intuitive, and intriguing."

My second thought was, "I wish they had this for my area, and with more information than just movie preferences. it would really be nice to find 'my place' in town."

My third thought was, "Wow, I can even see how the cities compare—at a glance!"

Then, I realized, "Hmmm. I could use this to find a like-minded neighborhood to move to. I could pre-screen locations to house-hunt in...." And that's when it hit me, I was zip-code-profiling!

I had been deceived, falling into the trap of hasty judgement using specious information. This is way over-generalized data. There is no indication of the sample size. For example, if Netflicks only rents to one person in a zip code, that person's preferences define the entire zip code area. Besides, a zip code area may be a convenient boundary for data collection, but it is far too generalized to indicate yours or my personal preferences, to any useful degree of certainty. The snazzy presentation hides those facts and encourages me to over-generalize and segregate, based on frivolous movie preferences.

On the other hand, with care I can delicately determine some relevant generalizations about regions. For example, San Fransisco's favorite movie was Milk. Not surprising considering it's subject. Dallas couldn't care less, with most of the city not even watching it (according to this data, however skewed). New York was really conflicted about it.

There is certainly entertainment value in looking over the site. Just take it with a shaker of salt (not just a grain). I still wish they had data for the Wasatch Front!