Evansville "Shots Reported" Data -- Part III
So far we’ve explored the data with a few charts and addressed the missing data issue. In this post, we’re going to look at the problem of clustering. That is, we’re going to see how often a single gun shot results in multiple phone calls.
To explore this phenomena, the 335 days were subset to the five days where the most shots were reported. Those calls were aggregated by the minute so that there were 1440 min for each day. The five timelines were laid side by side to see if and where the clustering occurred.
Plotting the days over the top of one another really shows how “shots reported” are an afternoon, early evening phenomena. You can also see that there are many occasions where multiple calls are made and some in the same minute. Certainly gun shots could be occurring in different places at the same time, but there could also be many where neighbors are reporting the same gunshot. A finally inquiry would be to see in addition to time how close the reports are in proximity. Fortunately, the data also include address data so it could be the subject of a future post.