The LA Times recently discovered that “nearly 40% of the crimes reported in official LAPD summaries did not appear on the publicly available map at lapdcrimemaps.org,” which in any case prohibits republication of its data and disclaims all responsibility for its accuracy. What’s the use of that–radically incomplete data nobody will stand behind, and you’re not supposed to copy and share it?

The City of Milwaukee and its police department make a similar disclaimer for the Community Mapping and Analysis for Safety Strategies (COMPASS) system, as well as their district crime maps used by MPD at community meetings each month. I’d like to think the data they put out from these sources represents close to 100% of the actual crime reports, but there is no way to know. There is no information provided to the public about what MPD crime data includes or excludes, or how it is created and disseminated through municipal data systems. At the same time, a private company–SpotCrime.com–is logging crime incidents in Milwaukee within hours of them occuring. You can’t get crime data anywhere close to that fast from the City. 

The police department has its own private data system, apparently much improved since Chief Flynn’s arrival due to his strong commitment to the CompStat method. It’s been a time of change for MPD, particularly for how it generates and uses crime data. Things were not in great shape previously, on many levels. There has been a need to drain the swamp of no-bid crony IT contractors who have been getting paid to do nothing. At some point soon, some accounting and explaining is in order–what have we got, exactly?

Right now it’s not possible to assess the data we do get, because we don’t know what is supposed to be included, what the critera are for inclusion/exclusion, and when the data is supposed to be available or when and why it might change. 

The closest you can get to crime data at its source are the 911 dispatch logs that MPD is now making public here. How this eventually corresponds (or doesn’t correspond) to incident reports isn’t clear; the logs just tell you what squads are doing now. If they are responding to an armed robbery call, they may not actually locate the person making the call or find sufficient reason to report that an armed robbery did occur. Or, shooting incident reports may be released but a related homicide doesn’t enter the official data until a later time. You have to wait a bit for the official incident reports, and people who frequently use the City’s eNotify service know it turns out crime incidents days, weeks, and sometimes months after they occur. The mapping systems will usually have data that is current within a week or so, but it’s not clear how complete it is, or how much it may be modified over time. 

As for Spotcrime.com, assuming they aren’t making things up, how are they getting Milwaukee crime data, and why are they so much faster at releasing it than the City? Are they buying the data? Originally they seemed to be republished eNotify data, but now they seem to have a faster source. They are definitely using local news sources and posting 911 logs–which are often not crime-related–so are they treating a 911 call about an armed robbery the same as an official incident report for an armed robbery? I have put these questions to the owner of SpotCrime.com, Colin Drane, and I have received answers that don’t really answer these specific questions. Drane has mentioned several times that he is concerned about cities and police departments selling data. According to TechCrunch, his business gets 90% of its data by harvesting it free from charge, but he has never said he doesn’t pay for it himself.

Apart from the allure of selling data–which has to look good to the City–every police department has every incentive to massage its data, especially when public scrutiny and accountability pressures from elected officials is lacking. As more pressure for transparency builds in an age when everything is digital and networked online, we can count on plenty of bad actors to use a pretense of transparency as just another way to massage a message. It might be nice to presume good faith on the part of people whose work is supposed to be in the public interest, requires the public’s trust, and is paid for by John Q. Public. It might be nice, but if that sentiment and our current state of ignorance about crime data is all we’ve got, we’re a bunch of suckers.