You aren't far off of something that is operational now. I worked in EMS and we had built models to predict where to pre-position ambulances by looking at the

You aren't far off of something that is operational now. I worked in EMS and we had built models to predict where to pre-position ambulances by looking at the demographics, past calls, complaints etc. When we were working on this, and many companies still work on this, another company I ran into was building the same thing for predicting where crime would happen for law enforcement using similar models. Of course, what it seemed to do to me was prejudice officers when responding to certain areas and I think does more harm then good in the end. I am not against using predictive algorithms, but it has to be done in a way which doesn't create prejudice in the people using it. The next step the company building the law enforcement models was moving toward was to use facial recognition and license plate readers which many cities had been adding to then let them track "criminal element" movements. Their argument is it would remove bias and track the criminals, yet the fact is it would track anyone, criminal or not, that the algorithm tagged as suspect. Which IMO is a major violation of peoples rights. No clue where it stands today as I haven't been around any of that in years, but I know license plate readers and cameras are growing in most cities so my hunch is it or projects like it moved forward.