Edmonton Police Service (EPS) Crime Mapping tool now online

Back in June we learned that the Edmonton Police Service was planning to launch a new website that would enable citizens to find crime statistics for their neighbourhoods. This afternoon, the EPS Crime Mapping tool went online, and it does just that. You can search for stats on eight types of crimes in any neighbourhood across any time period since 2007. From the press release:

The new crime mapping tool will provide members and citizens with a better understanding of what is going on the neighbourhoods they work and live in.

I’ve been playing with the site today, and I like it. There are pros and cons, however.

How It Works

The first step is to agree to the disclaimer – more on that in a minute. Next, you pick the crimes you want statistics for. The eight types include assault, break and enter, homicide, robbery, sexual assaults, theft from vehicle, theft of vehicle, and theft over $5000. Third, you pick the neighbourhood – there are 357 listed in the system. Finally, you select the time period. There are some quick selections such as yesterday or the last 30 days, or you can enter any two dates. Click “Show Crimes” and your neighbourhood appears on the map, covered in colored dots to represent the reported crimes. Here’s what Oliver looks like for the last 30 days with all crime types selected:

There’s also a “View Statistics” tab above the map that will show you a table for the last three years broken down by month, with a graph below that.

The Good

There are some really good things about this site. First and foremost, the data is excellent. I’m glad that they included everything up-front, instead of doing a test release or something to start. Second, it’s built using Google Maps. This is a big win for EPS – it’s a stable technology that Google is continually making better, and I would guess that most Edmontonians are familiar with it. Third, it’s fast. Almost as soon as you click the button, your stats appear.

The Bad

There are two things about the site that I don’t like. First is the disclaimer – it’s too restrictive. These two points in particular are problematic:

While it is acceptable to pass the website link on to others in your community, you will not share the information found on the website with others other than with members of the Edmonton Police Service or other law enforcement agencies; and

You will only use this website and the information in it so you can inform yourself of, and participate in, this community policing initiative;

That effectively means you can’t do anything with the data. This is in direct contrast with what the press release would lead you to believe:

Providing our citizens with the real picture of neighbourhood crime is the first step in engaging them to do something about it. Members of the public will be better equipped with knowledge to work collaboratively with the EPS to reduce and prevent crime.

What’s the point of making the data available if you can’t do anything with it? Why can’t I blog about the crime stats in a particular neighbourhood? Or mash the crime stats up with some other data? I challenge the notion that simply being able to see the dots on a map equips me to do something about crime in my neighbourhood.

I’ve emailed the feedback address listed on the site asking about this, but I haven’t yet received a response.

The second bad thing about the site is that while it does make data available, it does so in an opaque and closed way. If Edmonton is going to become an open city (with respect to data), sites like the crime mapping tool need to provide information for multiple audiences. One is the average citizen who is happy to click around on the map. Another increasingly important audience is the creative professional who wants to do something with the data, and needs it in a machine-readable format such as a CSV or XML file.

The Undocumented API

The first thing I did after testing the site with my neighbourhood was poke around for clues about where the data comes from. It didn’t take long to realize that there’s a JSON web service behind the application. You can access it here. It’s probably not meant for public consumption, but it’s there and it works. I was able to throw some code together in about 30 minutes to get data out of the service. While it would still be good to have static data files available, the API largely negates the con I mentioned above. As it is unofficial however, who knows if it will remain active and working, so enjoy it while you can.

Final Thoughts

Overall I think the Crime Mapping tool is excellent. We need more applications and services like this, though with less restrictive terms/licensing and easier-to-access data. Kudos to EPS for building this, and let’s hope they improve it.

UPDATE: There are more details in this article. For instance, the tool apparently cost $20,000 to build, and is automatically updated each morning.

41 thoughts on “Edmonton Police Service (EPS) Crime Mapping tool now online

  1. Good stuff. I would prefer to be able to see crimes over the whole city rather than having to choose each neighborhood individually.

  2. They could do it if they grouped the pins at higher zoom levels… Google Maps kinda sucks for that though, I haven’t experienced the performance benefit with pin clustering on Google Maps that I have on Bing. I think people who have it running nicely on GMaps have written their own custom layer code.
    Anyway… I can’t get the damn webservice to work. I keep getting SOAPExceptions when I try to invoke them. Something about being unable to generate a temporary class, which is usually a permissions issue on the local machine, but I can’t seem to resolve it even after having given Full Control of temp to Everyone, so I’m wondering if it might be coming from their server not my box… All I wanted to do was scrape all the polygon definitions for those neighborhoods!

  3. The info is telling me that I need to lock up my garage tight. Lots of that in my area. Thankfully there are no markers in my area for fashion faux pas (that’s the tartan coloured dot).

    Good job EPS it is a good start to what could be an great sharing tool!

  4. I also sent an email to the feedback link as well, asking for permalinks. Found some interesting stats I wanted to share and couldn’t. I love the Ajax in the interface but it’s relatively bad UX without permalinks. No response yet either.

  5. An obvious feature that I would like to see is updates or alerts when crime stats occur in my selected areas.

    This seems like a valuable feature in helping residents be aware of what is happening in their neighbouhood.

    That said, I like it and it’s a good use of technology in an open community.

  6. Removing them would kind of defeat the purpose most people want out of this, which is to find out how ‘risky’ a certain area is. Solved crimes still happened, and it’s not exactly a huge comfort to the people they happened to.

  7. There are serious privacy concerns associated with this tool that make it unlikely you will ever get access to the “raw data” behind the website. It would essentially open up the EPS complaint database to anyone with the time to regularly download the data and translate the latitude and longitude back into a street address. While it might be “fun” to know why the police attended at your neighbour’s house two weeks ago, I think you’re letting yourself get carried away with your rhetoric about “open cities.” There are perfectly legitimate reasons why it might be better to limit the public’s ability to directly manipulate the data.

  8. That’s a good counter-point BR. I wouldn’t say it’s directly involved with the “‘open cities’ rhetoric”, which mainly involves allowing us to consume and mash up all the data that’s out there already for the public but in formats we can’t process or mash up or behind eulas or disclaimers or what not that prevent us from doing so. Most of us who push for data openness don’t want any more info than you can already see on that screen, just the ability to overlay other data on top of it and that sort of thing.
    If you go to maps.edmonton.ca (and potentially jump through the hoops you might have to jump through to get it working) you’ll find that there is an amazing plethora of geolocated data publicly available to all citizens, but that we can’t pull any of it out into a nicer interface or combine it the way we’d like or search it the way we’d like or those kinds of things. That’s the kind of stuff that fuels the push behind openness as far as I’m concerned.

  9. BR – What Grant said 🙂

    I don’t necessarily want the crime data by lat/long, though I should point out that it’s already available through the API I mentioned above. More generalized statistics would be fine, similar to what EveryBlock provides: http://nyc.everyblock.com/crime/

    I don’t think there’s anything “fun” about crime data. I do think, however, that putting that data into the hands of citizens is a good thing. They’ll find innovative ways to use it. Sure there’s a risk, but I believe the potential benefits far outweigh any negatives.

  10. Another feature that would be useful is if we could choose multiple neighbourhoods. I am not saying that we need to see the entire city, but neighbourhoods that surround mine would be a good start. I am sure there are people that live on the boundary of a neighbourhood and really don’t want to have to search twice to see crimes in their area.

    I guess I am saying that searching by neighbourhood is too restrictive.

  11. Having the open data is great, but it seems odd that it is limited to a neighbourhood-by-neighbourhood search. I live at the corner of a neighbourhood, so to look up the crimes that are near me, I have to search four neighbourhoods separately.

    It’d be far more useful to just have the data overlayed on a city-wide map, and allow people to zoom in to see details. It’d be quite interesting to view this data for the city as a whole – it makes crime “hot spots” much more obvious.

  12. can’t wait till someone mushes it up with mls to watch trending house markets vs crime rates…or maybe a subjectively humourous site that tracks geotagged images across various image sites based on time theywere posted as posible witnesses…or perpetrators to create dynamic “wanted” or “have you seen this person”…maybe even with the ets bus schedule in relation to the “safest” bus routes on your way to the closest tim hortons?

  13. The possibilities are endless Mike 🙂

    Paul & George: according their help page, the restriction to a single neighbourhood is to “maintain an acceptable level of performance for the website”.

    I agree it would be useful to see more than one at a time.

  14. “it makes crime “hot spots” much more obvious”

    I almost think this is deliberate. Seeing the whole city at once would be kinda scary and demoralizing for people I think. This way you get the info, but not the relativity (eg, sure there’s dots in my neighborhood and I can see not to park my car on a certain block, but I don’t know if this is ‘lots’ of crime or not, because that’s all relative)

  15. Oakland has a crime map that asks for citizen input. It works by having a click through for each crime that accesses a comments page that provides a “tips” comments field plus the name of the case officer, a contact number and any relevant details. I will post the URL when I am back in front of my computer.

  16. lol

    I really should get my thoughts together in one comment, but hey..

    Being able to see what neighbourhoods area’s are would be very useful. For example, I have no idea what the neighbourhood north of downtown is called…

  17. Yeah, definitely a performance thing. I was admiring their asynchronous pin loader with progress bar actually, so they’ve obviously put some thought into making sure it’s snappy. I guess that brings us back to the desire to do pin clustering for performance reasons. I can’t recall why exactly I didn’t see a performance benefit when I did this in GMaps the way I saw it when I did it in Bing… I think it might have been particular to the situation I had; as I recall it was fine at higher zoom levels but unusable at zoom, and the difference was that Bing doesn’t seem to incur much of a performance hit if you add a pin to the map unless it’s actually visible in the map’s bounding box, but with GMaps it seemed to make almost no difference to the performance penalty whether or not the pin was visible. (I just can’t remember why I had to load all the pins onto the map at zoom instead of hooking into the map scroll event and loading them dynamically… whatever. I really should have taken notes. Maybe I did, somewhere. I’ll look that up one day when I’m back at my other office; I’ve been offsite for months.)
    Anyway, we could try it, if you really want to make 350 or whatever separate web service calls to load the pins, but I’m pretty sure that’d be a clear violation of their disclaimer.

  18. @Rob Davy, check out maps.edmonton.ca and it’ll show you the neighborhood name as you mouse over the area. Just sucks that it’s so clunky. I’ll put together something better…
    Actually Mack, I’m starting to feel like I should snowball up my ETS project to be just one feature of a general map mashup thing. Seems like we get more data all the time.

  19. I’ve been making really good progress, so should have something soon. Thing is, I’m not building an open source tool or anything like that. So there will be lots of opportunities to help out, but likly not with the code!

  20. Grant: the answer is maybe. I don’t really have any monetization plans at the moment, but I guess you could say I’m reserving the right to explore my options in the future.

  21. You need to show the whole City…or “ALL” on the map, it’s to difficult trying to figure out the names of areas unless you know every area which I don’t.

  22. Show the whole City on one map.
    It seems Government and Police are always trying to get away from showing the “WHOLE PICURE” or “STACKING THE DECK”…”IT ALWAYS GIVES THE IMMPRESSION THAT THE POLICE ARE HIDING THE WHOLE TRUTH”.



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s