Data Devolution: Topography of Crime

by Katia Savchuk

Next time my mom berates me for taking the bus late at night, I’ll be able to back up my claim that that particular part of downtown is totally safe at two in the morning. That’s thanks to SF Crimespotting, a new site that superimposes detailed crime data from police records onto an accessible and pretty mapping interface.

Built by Stamen, a design studio specializing in data visualization, SF Crimespotting – like its Oakland-focused predecessor – maps crimes in the city to street level with color-coded dots representing an array of transgressions. The snazziest part is that it lets you limit your search by time period (month, week, day or subsection thereof) or by broader intervals like “commute,” “dark” (nearest hour), and “swing shift.” They’re also working on RSS feeds that send alerts when crimes occur in areas you specify.

Besides the tool’s potential impacts on the real-estate market, crime prosecution and agoraphobia rates, Crimespotting is cool in that it recognizes the connection between space and the social. Most social analysis mistakenly considers the environment to be a neutral backdrop to interactions and ignores the interplay between people and space.

For one, representing information geographically can lead to some unexpected (and aesthetically pleasing) correlations. The image below, for example, shows trees, cabs and crime.

Source: Shawn Allen

SF Crimespotting can also serve as a myth buster (there were a lot more crimes around that downtown bus stop when it was light out than when it was dark!) that can not only calm my mom’s fears, but also potentially reshape our mental maps of the city. Realistically, though, most people won’t reference the site enough for this objective information to substantially alter their views of neighborhoods, which have a lot to do with personal experience and emotional valence.

The folks at Stamen, who undertook the project independently of the city or police department, are mappers with a mission:
We’ve found ourselves frustrated by the proprietary systems and long disclaimers that ultimately limit information available to the public. As citizens we have a right to public information. A clear understanding of our environment is essential to an informed citizenry.
Crimespotting democratizes access to geo-specific data, making it possible for anyone to get information about their environment. They make it even easier to share information by allowing you to pass on links to a particular view of the map. Decentralizing geographic data not only enables anyone to be a social scientist, but is critical for strengthening local decision-making and community planning efforts.


  1. Interesting! This is a unique way of looking at relationships between the social and spatial. I'm not an environmental determinist, but it makes sense that looking at actions without considering their surroundings gives an incomplete picture. Thanks for bringing these things up.

    P.S. At first I was confused about the lack of trees in Golden Gate Park, but I see now they're just showing trees monitored by Friends of the Urban Forest. Even without all of them represented, trees seem to outnumber crimes in most sections of the city. And it looks like crime follows street lines.

  2. Just came across an article that relates:

    From the abstract ... The authors investigate whether the percentage of green space in people’s living environment affects their feelings of social safety positively or negatively. Contrary to the common image of green space as a dangerous hiding place for criminal activity which causes feelings of insecurity, the results suggest that green space generally enhances feelings of social safety. The results also suggest, however, that green space in the most urban areas is a matter of concern with respect to social safety.

    The full article requires a subscription, but I'd like to read it to find out how they define "the most urban".


    Another interesting site detailing reported crimes in a particular area, quite useful and interesting for cross-referencing data.

  4. An East Coast version lets you pull up crime stats from local police precincts.

  5. Here are a few maps that show the exact same data, but in a different way and with different tools (RSS and KML feeds, an API, and mobile versions).

    San Francisco Crime:,%20CA

    Chicago Crime:,%20IL

  6. Thanks for all the links and insights! It's interesting that crimes seem to take place much more on the most populated street networks than in large urban parks, like Golden Gate Park or the Presidio. What does that say for the "eyes on the street" theory of safety? It would be interesting to see this based on the type of crime.

  7. One of the most basic axioms in geography, alright : The Nearness Principle.
    The closer elements or people are to each other, the more they will relate. Crowding reduces the distance between a knife and a victim's flesh.
    Makes a lot of sense.


Note: Only a member of this blog may post a comment.