While working on my dissertation, I've been thinking a lot about the role of CCTV cameras, and other forms of electronic surveillance. One especially touchy subject are red light cameras. While some research suggests they do have an impact on crashes at major intersections many opponents suggest they are simply money-making tools for cities. Indeed, citizens of Chicago are about to win a $39 million dollar settlement based on violations made by the city.
As a criminologist, I was curious about the distribution of violations across red light cameras. One topic I'm particularly interested in, is what we call the "law of crime concentration." What many scholars in criminology have found is that crime tends to occur disproportionately in a small number of places. For instance, my own work in Detroit finds that about 80% of shootings occur at about 3% of street blocks. This idea isn't new however and we've even coined a name for it - the "Pareto principle" or the "80/20" rule.
So out of curiosity I went to the Chicago Open Data website and used information on their red light cameras for the entire year of 2016. I obtained information about each camera and the number of violations per day. I then aggregated them by camera and sorted the cumulative number of cameras and violations descending. What I saw was pretty impressive: A whopping 25% of all red light camera violations were accounted for by 13 cameras - over 163,000 violations in total. If we assume that the average violation costs between $35 and $50, these 13 cameras are bringing in between 5.5 to 9 million dollars.
The curve appears to follow something like a beta distribution - with about 50% of violations accounted for by 20% of cameras, and about 80% of violations accounted for by 50% of cameras. This makes sense, based on what we know about how phenomenon tend to cluster in a few areas. Just a few number of intersections have a lot of people going through them, and also likely making illegal turns or going through the lights. Which intersections were these?
Number 1, by far, was 4200 S. Cicero Avenue - accounting for just under 30,000 violations in 2016. Second and third place were won by 400 W Belmont and 2800 W Diversey. To a Chicagoan this probably isn't a big surprise because these are all major avenues. More traffic = more opportunities for tickets. If I had traffic volume data, it would be interesting to see what intersections have the most violations per-car-hour.
In closing, the Pareto principle is alive and well and can be observed in all kinds of contexts. In some cases it can reveal interesting patterns, and give us pause. In criminology, we often want to know what crime "hot spots" look like - specifically, what sort of features might be generating that crime? For instance, what does our winner, 4200 S Cicero, look like?