A propos of nothing in particular the chart above is of police-reported motor vehicle thefts in Chicago from April 2008 through the end of December 2009. I compiled the data out of simple curiosity and once I’d done it I thought I may as well post it.
The data is drawn from a very interesting database of crime statistics provided at the Chicago Tribune’s web site. I may use it some time to produce a graph of the number of homicides related to the number of homicide convictions which I strongly suspect will be very interesting.
I’m not sure what conclusion if any to draw from the data above. The only thing that occurred to me was that you’d almost think there was a quota!
Is the X-axis measurement in weeks?
I was looking for a weather related effect; wouldn’t you assume vehicle thefts go down in Winter, up in Summer?
BTW/ convictions versus crimes is a tricky stat, particularly with new technology. Some PDs report more than a 100% conviction rate for the year because they are convicting more criminals from old cases with DNA and similar technologies than crimes are being committed.
Yes, the X-axis is measured in weeks. I thought that trying to cram all of those dates into the available space would have made it less rather than more readable.
To be honest I thought there woujld be a lot more seasonality than there turned out to be. I can only speculate that there’s something else going on.
Not to suggest work, but I like the stacked year graphs that Calculated Risk does for railroad traffic, hotel occupancy, etc. Then you can see both the seasonality (if any) and the year-over-year trend.