Benjamin Mokotoff on making the challenging medium of crime data accessible and interesting to people
Benjamin Mokotoff (@bmokoto) and Brian Sobel (@sobelito) are the guys behind the "Are You Safe?" (@AreYouSafe) application ("app") for the iPhone, which provides people in several cities (Milwaukee, Atlanta, Washington D.C., Sacramento, Indianapolis) with a location-based "threat meter" based on local crime data. The interface can also be changed to read "personal defcon," "suggested posture" (from "relaxed" to "full paranoia") and "suggested beverage" (from "chai tea" to "quad. espresso"). By default the app responds to your (iPhone's) location, but you can enter any address in the city and see how it rates, safetywise. I tried it out and posted my thoughts on it here: Ask your iPhone: "Am I safe?" Benjamin was kind enough to provide a very quick and thoughtful response to my questions, which you can read below. -Dan Knauss
Dan: Who do you think are the most common users of your app? How are they using it?
Ben: While Apple does not provide us with any information about the users that download our app, anecdotal evidence suggests that our most common users are people that are new to the city, tourists, real estate agents, and house hunters. The younger generation seems to really like the simplicity, and novelty of the app. It's not uncommon to see someone pull up the app, and then see him/her hand it to someone else in the group so that they can enter their address...and so on.
We talked a bit about our user demographics in the slide deck that we recently presented at the O'Reilly Gov2.0 conference in DC. It can be found at the folllowing URL:
http://www.slideshare.net/sobelito/areyousafe-municipal-and-federal-data-extraction-mashup-and-visualization
Dan: Do you think people use crime stats more these days to negotiate lower prices on home sales, and if so, do you think that is ultimately a good thing if they do this with apps like "Are You Safe?" [This was an idea suggested by "The Bus Bandit" -- @thebusbandit.]
Ben: It's an interesting question and I'm not sure that we know the answer. If we were looking for a home, we would certainly take crime rates into consideration in addition to school quality, proximity to public transportation/retail, etc. We're not sure neighborhood safety is a "negotiating chip" but more a driver of home value. Is it a good thing? Yes and no. We firmly believe that transparency into crime data is a good thing. However, when crime drives down home prices the result is fewer tax dollars to fund the local police department; this in turn opens the door to more crime. Somewhat of a vicious cycle in our opinion.
Are You Safe DC was recently written up by Mark Welborn or UrbanTurf. According to Mark, there crime rates are not something that real estate agents in DC are allowed to share with their clients. His article is worth a read and might provide some additional insight. It can be found at: http://dc.urbanturf.com/articles/blog/are_you_safe_in_the_neighborhood_you_are_moving_into/1248
Dan: What are the best ways to use "Are You Safe?"--in what kind of scenarios is it most useful?
Ben: AreYouSafe is a portal, mash-up, whatever you want to call it of hyper-local (i.e. street address specific) ground truth data. Our products are both the iPhone application, and the heat-maps that we make available on the landing pages. In the app, we provide data in 3 forms to the user, (1) relative safety index based on quantity and severity of crimes per city grid point (2 block radius), (2) number of crimes broken out by type in that radius, and (3) closet crime to where the users is located (or to the manually entered address that was input).
We can't necessarily say what is "the best way", but we can share how we know people to have used it. We've heard from some users that they simply like to know the level of crime where they are or where they are going out in the evening. We've heard from business travelers that it's a useful tool when looking for a place to stay or a place to grab a bite to eat close to their hotel. We heard from a family that they had 2 iPhones with them when visiting Washington DC recently and that their kids loved seeing the meter change as they walked/cabbed around the city - somewhat of a novelty to them. In that novelty sense, we know some users use it as such: driving around trying to find the "red" areas.
As far as determing where it is most useful - that's up to the eye of the user. We built the app because there was an opportunity to utilize this data in an innovative way, not because we thought people should be using it as a magic 8 ball for decision making purposes. However, the feedback we have gotten is that users are using it to assist them in evaluating the historical safety of locations they frequent or plan to check out. Again, I think over time, stories from our users will tell us how it was "most useful" to them -- we look forward to hearing such.
Dan: Are there ways people are using it you didn't anticipate?
Ben: Above and beyond what we've described in the answer's to the previous questions - not really. Those responses pretty much covered it.
Dan: How did you come up with the Brady Street and Riverwest references for the ad text in the app store for the Milwaukee version? The way that was written, it appears to cater to people who do not know the city well, live downtown or outside of the city, and are wondering if these neighborhood areas are "safe." Is that the intention? What would you say to business owners or residents in those areas who resent being singled out as potentially "unsafe?" Or more broadly, what would you say to public officials and city residents who react negatively to your app and think it is just going to spread fear to people who don't have much knowledge or context for understanding an area? [Another set of ideas from @thebusbandit and also @creamcity, with my own spin on it from familiarity with how city residents often worry about negative safety perceptions.]
Ben: Fair questions, and we've certainly received feedback of this nature before, re: "Fear Mongering" and "Paranoia". We've also received feedback from users who are thrilled to finally have a way to visualize crime data in a mobile fashion. We look to pick out a couple of "representative" areas within each city for our app description and they are, as you put it, "singled out." We come up with a list of what we think are the most "recognizable" places in each city and then choose 2 or 3 to include. It is by no means our intention to create a state of fear with regard to these areas for our users. That being said, we do want to provide those users a portal into the mass of data that the cities we support provide, in what we would alternatively consider a "challenging medium" for consumption.
Dan: How exactly do you compute the safety/threat level? Can you share the secret formula?
Ben: While at this point we aren't comfortable sharing the exact formula, we will take it under consideration. We will have to weigh the transparency angle with the business trade secret angle, though here is insight onto the formula and how it is applied:
We use the same formula across every city we process. It was worked out using Washington DC as the driver, as we are intimately familiar with the city. We break a city down into two-block radius "grid-points" and then calculate a score for each based on the severity and quantity of incidents in that radius. Each crime-type is weighted, for example, sexual-assault rates higher than car theft, and homicide's "score" is 40*(number of occurrences)^3 so that one chance homicide in an area won't skew the rating significantly, but a pattern of homicides in an area will. We do this for the entire city. Knowing that the more densely populated an area is the more crime there will be, we use the census data to "level the playing field" if you will. We do this by finding the closest "census group" to each "grid-point", and then multiply the score from above to the people per square mile in that "census group". After acquiring the new census affected score for each "grid-point", we apply a curve to break all the score into the 10 safety levels one see's in the "threat meter".
Slides 7 and 8 in the SlideShare presentation referenced above (http://www.slideshare.net/sobelito/areyousafe-municipal-and-federal-data-extraction-mashup-and-visualization) do a nice job outlining process we used to create the app. The video from the conference should be available soon which will also shed some additional light on the process.
Dan: Are there any planned new features for future versions, like more data for better trend analysis? Something simple you could do with the existing data set is indicate the direction(s) of the nearest inrease or decrease in crime concentration.
Ben: Our focus, at least at the moment, it breadth over depth. We'd like to roll AreYouSafe out to a few more cities that openly share their data before driving deeper from a functional perspective.
We've had a lot of users suggest new features and we have a list of "enhancements" that are on the horizon. The most frequent request have been for safety ratings relative to the time of day and more than one record with regard to the closest crime (the option to scroll through, for example, the 10 closest crimes). Your suggestion regarding the direction of the nearest increase/decrease in crime concentration is a good one and might be a nice opportunity to leverage the 3GS's compass capabilities. Thank you!
Dan: Some cities, like Milwaukee, put out location-based crime data on a monthly or even more frequent basis. MPD dispatch logs are also online with a 90 minute delay. Have you thought about accessing that kind of data for something closer to a real-time analysis?
Ben: We've absolutely discussed the frequency of data updates internally and know that there's an opportunity to make more frequent updates. It is something we are actively looking at. Apple has recently introduced a capability called push notifications that would make near-real time updates possible. Our team will be exploring this in future as well. While some cities do put out location-based crime data on a monthly or even more frequent basis others don't put out anything at all. As city governments become more transparent, developers like us - and all the others out there, will have the capability to aggregate the data and present it to our users in more meaningful and timely ways.
Dan: Do you consider "Are You Safe" a "Gov 2.0" app? Compared to apps like those offered by SeeClickFix and CitySourced (where people report non-emergency issues, like potholes), it doesn't offer a means of "civic engagement" or any interactivity with other users. Are there ways that might be relevant to add, like letting users post notes to locations and shared these across their city network? Is that kind of collaboration a necessary part of a full blown Gov 2.0 type of app?
Ben: There are a lot of parts to the Gov2.0 movement: Citizen collaboration, transparency, data availability, use of social media tools, use of open source software, making goverment more efficienct throigh use of 2.0 technologies, etc. It is our opinion that one can participate in Gov2.0 without hitting every buzz word. We all have parts to play, anyone can be one or a few cogs of the Gov2.0 machine, and in that part contribute ot the movement. We think AreYouSafe falls into the solution provider, data consumer, innovationator categories. These are essential parts of the Gov2.0 ecosystem, just as much as making the data available.
As side note, we think the"civic engagement" attribute you referenced is a noble Gov2.0 angle, and the work that Ben Berkowitz of SeeClickFix, and the whole 311 app movement is exciting, and will really benefit citizens by making government more nimble, responsive, and cost-effective.
Trackback(0)
TrackBack URI for this entryComments (0)
Subscribe to this comment's feedShow/hide comments
Write comment
Show/hide comment form
