Friday, January 22, 2010

Good Ideas for Cycling Laws

The Greater Greater Washington blog is a great resource for people interested in increasing the availability and safety of alternate forms of transportation. Today they have a really good post on 12 Ways Our Region Could Reform Bicycling Laws. If you're even a casual cyclist in Maryland, you should read this article. I know that since I've been reading that blog I've learned that not only are the roads less safe for cyclists than I thought, but that generally cyclists don't have much of a legal leg to stand on in Maryland if they do get injured (or even killed!) by a car.

Sunday, January 10, 2010

Baltimore 2009 Homicides - Everyone CALM DOWN

So the opinion in this town seems to be that Baltimore had 238 homicides in 2009 compared to an official 234 in 2008. Nationally homicides decreased in 2009. DC was way down this year. Quick, everyone panic!!

Bulls&%*.

In 2009 we had 14 prior year homicides. That means that 224 people died of injuries sustained in violent incidents in the city of Baltimore in 2009. 4 of those 14 prior year assaults occurred in 2008. Baltimore logged around 140 fewer non-fatal shootings in 2009 than 2008. As much as critics and "The Wire" fans would hate to admit, the homicide rate is lower and this city is safer than it has been in a generation.

The Baltimore Sun ran a very good series of graphics on what homicides and shootings looked like in 2009. The Sun shows that in 2009 there were around 450 non-fatal shooting in the city. While you might be like "DAMN!", if you look at the graphic you'd realize that only 2 years ago there were 652 non-fatal shootings. That's an incredible drop over a period of only 2 years. This whole "fight against bad guys with guys" that Bealefield is always on about seems to be making a difference. Frankly, given the level of non-fatal shootings, I'm shocked we managed to make it up to 238 for the year.

Actually, that's where homicides due to prior year assaults comes in. As I stated in the intro, Baltimore racked up 14 CY2009 homicides from prior year assaults. That's a lot. Per my research of the Murder Ink archives, we had 5 prior year homicides in '07 and 6 in '08. 14 is a big deal and it skews the real measure of how deadly the city was last year. In 2007, when the city record an official 282 homicides, 5 were a result of prior year result, which is about 1.8% of homicides. In 2009, we have 238 official homicides and 14 prior year assaults, which is 5.9% of homicides.

What this calls out for, no, cries out for, is someone to track homicides based on the date of the assault! Thankfully, I am willing to do this. I've carefully researched Anna Ditkoff's Murder Ink and the BaltimoreCrime blog posts. I've spent time auditing my numbers. Lots of time. My girlfriend has been frustrated by the time I've spent on this. But I've got your answer.

Discounting prior-year assaults, there were 279 murders in 2007, 232 murders in 2008, and 224 murders in 2009 as of the first week of January, 2010.

"But," you might say, "I don't understand! You said there were 282 murders in 2007 with 5 as prior-year assaults, so why is your total 279?" The answer is that I'm keeping a master spreadsheet. So if someone died in 2009 due to an attack in 2007, I go back and add the murder to the date in 2007 when the victim was attacked.

This master spreadsheet allows me to track a number that I call "Trailing Twelve Months Homicides". This statistic does not appear to be tracked anywhere else, but I think it matters. Restarting the homicide count every January doesn't show you how you're really doing. If you start counting every January, by March you're comparing 33 murders to 30 and coming up with a 10% increase in homicides when really it was just a couple bad weekends. There aren't a lot of homicides, really, so the data is really noisy and needs to be considered over at least a 12-month time period. So I think that Trailing Twelve Month Murders really makes sense.

So, without further ado, here's the plot of Trailing Twelve Month Murders from January 1st 2008 to approximately the present:

(Click on the chart above if you want to a better view of it, Blogger doesn't do a real good job with graphics sometimes)

Anyway, look at that graph for a minute. It shows that two short years ago the Baltimore murder rate was around 275 for the past 12 months. Today it's sitting much closer to 225. Back in mid-2008 we dipped below 215 per 12-month period before jumping back up to around 230. Most of '09 held steady around 230 (the 30-day moving average never got close to 250 for all of 2009). During the fourth quarter of 2009 the murder rate started slowly sliding downward. By the end of the year Baltimore was doing better than we had since 2008.

Why didn't you know this already? Because around Thanksgiving the ME's office started owning up to all the people that died this year who should be classified as homicides but were actually assaulted in a previous year. Ok, maybe it wasn't the ME's office. I actually have no idea who actually "owns up" to those statistics, but the point is that it hit us hard in December. Between Thanksgiving and the end of the year we added 8 prior-year assaults to the '09 homicide tally. So right as everyone started paying attention to the annual total, it jumped for an external reason. Prior-year homicides took December from a total of 19 homicides (about average for 2009) to 25 (a pretty bad month for 2008-09).

So the moral of this story is that Baltimore is on the right track. We're holding steady against last year (the best year for homicides in two decades) and even if we hold steady on new attacks, we're likely to drop the total. Non-fatal shootings are basically the feeder statistic for prior-year assault homicides and those have been declining, so I'd expect fewer of those next year.

Anyway, I may well dig up some more statistics about murders, but let me end this post with a prediction: I predict that 2010 will have no more than 215 homicides. When you add in the prior-year assaults, I'll take 220 as an upper bound for the official homicide tally in 2010.


Summary graphs for the year:

Baltimore Murders 2007-09



Finally, for the big nerds out there, here's the summary of homicides resulting from prior year assaults:

2007 - 5 (1991, 1997, 2x 1999, 2005)
2008 - 6 (1985, 1995, 1996, 2000, 2006, 2007)
2009 - 14 (1988, 1994, 1997, 1998, 1999, 2002, 2003, 2005, 2006, 2007, 4x 2008)


Also, here's a link to my static XML murder charts.