Google Maps and Crime Statistics.

When Google was in town recently, one of the applications they highlighted was a map someone had created which displayed crime information in the city of Baltimore, Maryland. Since then, crime maps of various cities seem to be getting increased exposure. Therefore, here is the story of one I created for Grand Rapids.

Over the past ten years, my car has been broken into four times. The first two were simple take-coins-from-the-cup-holder incidents. The last two involved broken windows and stolen stereos. After replacing the second stereo I started looking around to see if other Grand Rapids folks had suffered similar indignities. There was no single source of local crime information available at the time. So, in the summer of 2006, I decided to do something about it.

In 2005, journalist Adrian Holovaty created the Chicago Crime Map. This map had everything I wanted, and was getting a lot of positive attention. After a little digging, I found a tutorial, downloaded a Google code library, and in about an hour I had my first map.

Creating the Google Maps mash-up was easy. Finding information with which to populate it took a little more work. The GR Police Department publishes yearly statistics, but at the time, nothing with the level of data granularity I needed. So I turned to the media. Crimes are reported all the time. Grand Rapids has several news websites. Finding things to post would be easy.

Unfortunately, not as easy as I had hoped. At the start of this project, only WOODTV had information accessible, archived, up-to-date, and information-rich. The rest of the local news outlets were too difficult to search, too slow to update, or simply didn’t cover enough local news. So WOODTV became my information source for the map.

One of the unspoken rules of journalism is “If it bleeds, it leads.” The more lurid the crime, the more attention it gets. A drive-by shooting gets in-depth coverage (as it should). A bicycle stolen from a back porch goes by unnoticed. Given the nature of this beast, the information I most wanted was not available from traditional media outlets. I used what I could find, so the crime map leans heavily toward violent crimes. Not a misdemeanor to be seen.

PRESENT:

This is the process I use for updating the map: I visit the local news websites a couple of times a day. When an incident is reported, I grab the direct link to the story, then go to maps.google.com and enter the address (if provided), or hunt around until I find the location (“near the intersection of X and Y streets”) From this I can retrieve the zip code, city quadrant, and latitude/longitude information, all of which I then enter into the XML file which contains all of the information for the crime map.

In raw code, each incident looks like this:

<incident>

        <type>Homicide</type>

        <date>2006.08.07</date>

        <street>1100 Crime Street SE</street>

        <city>Grand Rapids, MI 49506</city>

        <quadrant>SE</quadrant>

        <latitude>42.953</latitude>

        <longitude>-85.641</longitude>

        <notes>Joe Smith shot and killed in his house</notes>

        <news_source>

               <n>http://woodtv.com/Global/story.asp?S=xxxxxx</n>

        </news_source>

</incident>

Once completed, I upload the XML file to the website and voila! Up-to-date crime statistics. Or as close as I have been able to get so far.

FUTURE:

With this as a base, there are many directions I could go with this map. The most useful would be to plug directly into the database at the Grand Rapids Police Department. That way every reported crime would be visible, and given the information density, the map would become much more useful.

I would like to see crimes broken down by time of day, as well as date and location.

I would like to add a timeline with an auto-play feature, so users can discover patterns in the locations and types on any given day.

Example: Look at all the crimes committed over a 24-hour period. Time-stamp them. Use the location to back-track and perhaps discover the origin of a particular spree. This would have been useful the second time my stereo was stolen, as this was part of a rash of car break-ins, at least a dozen of which happened on the same block in a week.

Perhaps most important, I want to have details of minor property crimes available. For every murder in Grand Rapids, there are probably a couple of hundred burglaries. I would think that people would get more use out knowing about the quiet, minor stuff, than seeing major news stories duplicated.

I would also like to convert the page over to being run by a blogging tool of some kind, so people can comment on the events of the day.

The crimes which are reported by the media are the smallest tip of the iceberg of the actual incidents that happen in a city. In 2007 I placed around a hundred incidents on the map. In that year there were roughly 15,000 crimes – felonies and misdemeanors - committed in the greater Grand Rapids area.

CURRENT AND POTENTIAL REACTIONS

Not long after I created the map I began to get referral hits from a local church. A little while later, I began to get hits from the Urban Planet forums. And shortly thereafter I discovered myself on the front page of Google when searching for “grand rapids crime”, or any variation thereof. So there is an interest in this kind of information being made publicly available.

Recently, I contacted Adrian Holovaty and asked him how people had reacted to the creation of the Chicago Crime Map. He said that everyone who had seen it – citizens, police, and local government officials, were happy to have it available, and somewhat surprised that nothing like this had been created previously. The only criticism was the lack of specific crime details. Mr. Holovaty recommended that I add a disclaimer to the effect that the data presented in the map is only as accurate as the sources from which it is gathered. Inaccurate data coming in equals inaccurate data going out.

This issue brought into focus another issue: the right to privacy of the victims of residential crimes. For most such incidents, the news sources reference the nearest crossroads, rather than specific addresses. I try to place the markers on the map appropriately, or in the middle of the block if a crime happened in “the 1600 block of SomeStreet”. For public space crimes – bank robberies, and the like – specific addresses are generally offered, and I place the markers as close to dead-on as I can.

In my opinion, more information can only be a good thing. The more we know about the places we live, the safer we are. The potential exists for businesses to suffer if people know that crimes have been committed nearby, but in the long view being well-informed is more important,

CODA

The crime map is now more than a year and a half old. Between 250 and 300 people view the page every month, and those numbers seem to be (slowly) growing. Soon I will begin a conversation with the GRPD regarding the use of their information in my project.

I for one would have like to have seen the statistics on stolen car stereos before I decide to leave my car unattended on a particular curb. If I can prevent one car stereo from being ripped out of a dashboard, I will consider this project to be a success.

OTHER MAPS

The following is a list of other crime map mash-ups in other cities in the US.

Chicago: http://chicago.everyblock.com/crime/ (the first widely-known crime map mashup, created by Adrian Holovaty)

New England: http://www.incidentlog.com/

New York: http://www.nytimes.com/packages/html/nyregion/20060428_HOMICIDE_MAP.html

Los Angeles: http://www.latimes.com/news/local/crime/homicidemap/

Oakland County: http://oakland.crimespotting.org

Large, detailed list of current crime maps: http://journalistopia.com/2007/09/06/journalistopias-online-crime-maps-directory/