Biased crime data and predictive policing

Mara Hvistendahl, (Science magazine 30 Sept 2016) had a great review article that I just got to read carefully.  It’s a good review of how big data and “predictive policing” (those are definitely scare quotes!) are promising a lot, perhaps clarifying some things but also probably perpetuating existing biases.  As someone watching the Oakland Police Department collect data and hoping that sharing it with Oakland citizens is a Good Idea, it seems to contain many cautionary tales.

I’ll be curious as to your reactions to the Science article, but since it’s trapped behind a paywall I’ve also captured the main text here  .

Another study of PredPol, using “synthetic” data

Thanks to EastBayExpress’s Darwin BondGraham tweet, I just saw this report from USAToday analyzing drug crime data from 2010. (Their underlying paper was published in Significance, October, 2016.)  Can we find correlation in our real OakCrimeData set?!

Leave a Reply

Your email address will not be published. Required fields are marked *