Understanding St. Louis: Homicides 2005-2012

Tweet about this on TwitterShare on Facebook1Share on Reddit0Print this pageEmail this to someone

City of St. Louis Homicides by Neighborhood 2005-2012

The graphic above is meant to nominally mirror the one created by the NY Times and titled "A Chicago Divided by Killings". There were 506 killings in Chicago in 2012, a 17% increase over the previous year and the second highest annual total since 2003. Twenty-first Ward Alderman Antonio French Tweeted a link to the Times' graphic and stated that it would interesting to see a similar map of St. Louis. I agreed and figured I'd give it a shot.

The premise of the Chicago item was that killings there followed that city's racial divides. The St. Louis map would seem to show very much the same. Race, income and educational attainment are closely aligned in neighborhoods across the city. The divide is so great, that in the "near homicides" neighborhoods (the 20 in which there were 20+ homicides over this period), the average homicide rate is estimated at 57 per 100,000 residents, while the rate in the "not near homicides" neighborhoods (the 20 that saw 0 or 1 total homicides) is estimated at 1 per 100,000 residents. There are more than two St. Louis's, but there are certainly two incredible extremes in the city. Still, the most stark numbers may be the victims. Of the 567 homicides from 2008 to 2011, for which the race of the victim is available in the SLMPD annual reports, 502 are listed as black, while 64 were white. Over that period, 89% of those killed in the city were black. In a city that's very nearly 50/50 black/white, those 64 homicides would give an annual murder rate of ~10/100,000 for white residents and ~78/100,000 black residents.

Notes: Homicide data is taken from end-of-the-year reports available on the Metropolitan Police Department website and cover January 2005 through November 2012. Racial composition of neighborhoods uses 2010 Census data and economic and education measures rely on numbers from the American Community Survey. The measure of "near homicides" includes the 20 neighborhoods that saw 20+ homicides over the dates covered. The measure of "not near homicides" includes the 20 neighborhoods that saw zero or one homicide over the same period.

*This graphic was updated on 1/7/2013 to reflect corrections in several neighborhood homicide counts. The largest changes occurred in the neighborhoods of Clifton Heights (12 to 2), Ellendale (8 to 0) and O'Fallon (29 to 45), as well as several city parks. The errors were the result of neighborhood name changes that resulted in misaligned tabulations. The corrections have enhanced the disparities apparent in the map, most significantly increasing the homicide rate per 100,000 black city residents from 53 to 57 and decreasing that of white residents from 2 to 1 per 100,000. The old map and revised verision can be seen together here.

All data used to produce the graphic and associated numbers above can be found here.

Tweet about this on TwitterShare on Facebook1Share on Reddit0Print this pageEmail this to someone
  • Pingback: roof repair()

  • Pingback: wedding planners()

  • Pingback: hop over to here()

  • Pingback: orospu()

  • Pingback: zdporn.com()

  • Pingback: GVK Biosciences()

  • Pingback: sitesell business()

  • Pingback: syarat cpns 2017()

  • Jay

    This might seem nit-picky, but not every neighborhood has a racial majority (meaning 50% or more of the population). For example, the Mt. Pleasant neighborhood has roughly 4,400 residents and less than 2,100 are black. The map shows this as a majority black neighborhood. It’s more than the 1,900 white residents, but not a majority. Still an interesting map, thanks for sharing.

    • Alex Ihnen

      That’s a fair point. While not meaning to dismiss other races, I wanted to graphic to be easy to understand, while still reading true. For what it’s worth, Mt. Pleasant was 59% white/32% black in 2000 and 43% white/47% black in 2010. It’s likely that it’s a majority black neighborhood today.

  • Can you explain your educational attainment categories a bit?

    • Alex Ihnen

      They could be more descriptive, but they use the American Community Survey data. Categories such as “some high school” are left out for simplicity. I just used the same categories as the NY Times Chicago map. When I get more time I’ll put a spreadsheet online to share the data used in one place.

      • Thanks. Was just wondering if “high schools” meant both “some high school” and “high school diploma” or just the latter. Same for “Bachelor +” There was some discussion on Reddit and at least one person was coming to the conclusion that only 29-41 percent in the city had graduated high school. There are obviously some missing people, presumably “some high school” and “some college” folks.

        • Alex Ihnen

          Right. You can’t add the numbers provided in the graphic to get high school grad completion rates. When I can get the spreadsheet together it will provide more useable stats.

  • Daniel Layton

    Excellent map. Have you ever done a map or animation that shows the changing racial composition of St. Louis neighborhoods at 10 year intervals with census data? I was surprised to learn that my neighborhood (Tower Grove East) was majority white and was thinking such a graphic would be interesting to see.

    • Alex Ihnen

      There is one (though you can’t zoom in for much detail) – check out Colin Gordon’s Mapping Decline interactive site here: https://nextstl.com/themedia/mapping-decline-online-st-louis-and-the-american-city. His book by the same name is a must-read. I lived in FPSE/The Grove from ’06-’11 and that neighborhood had a large white majority until the 1980’s (I’m told), then has had a black majority and is now roughly 50/50 and is trending back to a white neighborhood.

  • abernajb

    Please cite your reports. I can’t seem to find annual reports that show murder by neighborhood.

    • abernajb

      Ok, so I found the reports:


      With the Annual reports showing under the December timeframe, but your map confuses me. For instance, it seems to show 12 murders in Clifton Heights, but looking at the individual reports from the above site, I see only two murders, one in 2009, and one in 2012.

      • Alex Ihnen

        Thanks for catching this. Some of the numbers didn’t seem right – as noted by others (Compton Heights, O’Fallon Park, etc.). Looking back through the spreadsheets, I found that statistics from 2005-2007 were not accurate due to neighborhood name changes. This meant that some homicides were misattributed as columns didn’t align properly. The graphic above should be accurate. While some neighborhood numbers did change, most did not and the overall map changed little. Thanks again for finding this. The old graphic can be seen alongside the new one here: http://www.flickr.com/photos/nextstl/8357552032/in/photostream

  • M Powers

    17 murders in O’Fallon Park (the park itself) seems particularly high. Same with there being 5 murders in Carondelet Park. There may be an error in SLMPD data here when the neighborhood name and the park name are the same. Also, three murders in the Botanical Gardens?

    • kcdoc

      I believe you are reading the graph incorrect. It shows one homicide in O’fallon park, 2 in Carondelet and none in the Botantical Gardens. Hope that helps.

      • Alex Ihnen

        An early version of the map showed what Michael points out. Numbers were transcribed incorrectly. The map was updated a couple hours after being posted and is now correct.

  • This is a very nice contribution, making points not unlike those I tried to make in a piece I wrote for nextSTL this time last year, “Yes, St. Louis is becoming a safer place” https://nextstl.com/urban-living/yes-st-louis-is-becoming-a-safer-place

    No one would be surprised if a geographical array of homicides coincided in striking ways with the greatest concentrations of poverty in the City of St. Louis. But many St. Louisans might be surprised to see, as nextSTL’s map reveals, how much of the city has an exceedingly low homicide rate. I haven’t checked nextSTL’s calculations of homicides rates in the neighborhoods it categorizes as “not near homicides”, including many with significant economic diversity, but 2 per 100,000 is less than one-half of the national average. What’s more, neighborhoods categorized as “near homicides” seldom get public credit for large sections, including parks, that are quiet and stable.

    Mapping lies at the heart of the steadily and increasingly pervasive “hot spot” police strategy being employed in St. Louis, one in which the incidence of violent or other targeted crime is geographically pinpointed both for historical periods and in real-time. Police resources then are systematically marshaled and deployed to micro areas experiencing higher concentrations of crime, with the nature and affects of police intervention constantly recorded and monitored — here by independent evaluators from University of Missouri-St. Louis’ celebrated criminology department as part of the Public Safety Partnership Mayor Slay organized in December 2011.

    The premise underlying such crime prevention strategies, as described by the brilliant criminologist Frank Zimring in his brilliant book, “The City that Became Safe,” is that crime can be seen as “contingent” and “situational” — the crime you are able to disrupt or defer today is not inevitable, it is not sure to reemerge, it is unlikely to be displaced to a nearby location, but instead may be prevented.

    Mapping can offer other valuable situational context, such as factoring in known conditions and risks specific to the targeted crime. What, for example, would nextSTL’s map and homicide rates look like if displayed homicides were limited to (or excluded) those occurring during late night (or work day or early evening) hours? How would the array and rates appear if they were limited to (or excluded) homicides known to be the product of gang violence, or family or intimate partner violence, or in which there is direct evidence of drug use among all known participants, or in which all known participants had prior criminal convictions, or exclusively involved men between the ages of 18 and 25.

    How might such dynamic mapping affect the public’s perception of safety and well being and help it organize its own conduct to reduce risks of criminal victimization? How might it inform and affect policy makers’ and police and other public safety strategic and tactical approach to law enforcement and crime prevention?

    This too is work being undertaken by St. Louis’ Public Safety Partnership.

    For decades, St. Louis’ homicides rates have been damnable and tragic, notwithstanding billions of dollars of investment in police and other public safety and neighborhood stabilization strategies. Rates remain unacceptably high — especially in neighborhoods with high concentrations of poverty.

    But St. Louisans should take heart in knowing, as nextSTL’s map reveals, that city neighborhood upon city neighborhood are and always have been safe, that broad measures of violent and property crime show a city that has experienced indisputably steady and dramatic reductions in crime in recent years, and that St. Louis has the potential, and I believe is poised, to experience breakthrough reductions of crime.

    • Alex Ihnen

      Great comments, Eddie. Thanks. It would be wonderful if the City/SLMPD would take up many of your points and questions and produce maps and information easily accessible to the public. Surely the number of homicides that could be categorized as random is very, very small. I agree with many of your points, and don’t think that highlighting how safe large swaths of the city have been and remain, somehow disregards the continues issues elsewhere (as some charge). That said, I think anyone’s immediate take away from the map likely isn’t how safe parts of the city are, but how there is a stark divide between “safe” and “unsafe” parts of the city. None of this is surprising. We know that crime/income (poverty)/education all correlate to crime (and in this case homicide).

  • megrot

    This is a great resource, thanks Alex.

  • onecity

    Great chart. Great distillation of the issues: economics and education.