Open Data for crime prevention: some food for thought from Brazil

Last October I was invited to the first Brazilian National Meeting on Open Data. One of the things I brought back is lots of thoughts about the potential of Open Data about crime.


A laudable and not negligible part of this potential is in crime prevention. Another, equally important part is how much complete Open Data about crime can help:

  • public administrators to plan and justify effective social policies
  • and citizens to know what’s really going on, beyond prejudices and stereotypes.

Luis Fernando Linch is the current manager of the Crime Statistics Division of the State Secretariat for Public Security of Rio Grande do Sul. A couple of years ago, Linch worked with others on a report about traffic accidents in the Rio Grande Do Sul state. The report showed in detail where and when most accidents happen, as well as the profiles of involved drivers, thus helping to recognize and plan the best actions for drivers education and traffic law enforcement.


During the 2011 Open Data National Meeting in Brasilia, Linch presented even more impressive data.

As in many other countries, most killers of women in Brazil are people very close to them, who somehow also “announced” their intention through previous attacks.

During his talk, Linch explained that 91.23% of the cases of murdered women had been preceded by other attacks (often by the same people who would eventually kill them), with an average interval between killing and the last aggression of 205 days: “We know that women attacked would die”.

As chilling as they sound, statements like this can be of huge help for social planners and citizens alike. When Open Data of this kind are available, they don’t generate fear (not if they’re explained properly, anyway). On the contrary, it becomes much easier both to take decisions and to have them accepted. In Linch’s words, “the way to fight this is to democratize information”. Today, however, this doesn’t happen because of, among other things, disconnected actions, programs and bad databases not connected. That’s lack of Open Data, if you ask me.

With number like those, for example, it may be much harder to criticize statements like “we need to hire counsellors and psychologists specially trained to support women who are victims of physical attacks, even if this means that remodeling the city center will have to wait one more year”.

Heck, were it really necessary, with more Open Data of the right kind, one may even go the cynical route and prove that it makes financial sense because of all the money you’d save in investigations, trials and detentions…


The idea makes even more sense if you think that it’s general. Linch noted that similar statistics may be made about children and elder citizens. Finally, Linch showed maps of some cities in which you could see by yourself that the neighborhoods with the most homicides very much overlap with those inhabited almost exclusively by people who earn from 50 to 100% of the Brazilian minimum wage. Now, if you had similar numbers (from official sources, not “feelings” or politicians press releases) for your town, how would you explain them, and what would you do about it?


Share

16 thoughts on “Open Data for crime prevention: some food for thought from Brazil

  1. Pingback: A few notes from the 2011 Brazilian ICT meeting | Stop

  2. Augusto Herrmann

    Hi, Marco!

    It’s nice to know that the Brazilian Open Data Meeting has been useful to you and has brought you some insights.

    As for Open Data about crime, specifically for data concerning violence against women, I think it’s more of a government database interoperability issue than open data itself. I mean, you’d generally want that the various levels of law enforcement and the judicial system could have integrated databases and share data.

    However, exposing this data to the general public would raise some very serious privacy concerns. What woman would want everyone of their neighbours and colleagues at work to know that she’s been assaulted? No one, I believe. That was the first thought that occurred to me as I watched Linch’s presentation.

    I think open data and government information systems interoperability are two issues that are very much related and usually go hand-in-hand, except when there are privacy or security issues involved. I think this is the case here. I don’t see a benefit in conflating these two related, but different concepts. So, in order to not raise some justified privacy concerns, I’d prefer to call this an interoperability case, or “shared government data”, if you will – which still is very much useful to society despite not being a total disclosure of data.

    Cheers,
    Augusto Herrmann

    1. marco Post author

      Hi Augusto,
      What woman would want everyone of their neighbours and colleagues at work to know that she’s been assaulted? No one, I believe

      I think the same. But there is no reason to publish names and addresses in cases like this. I never thought they should, and I suppose Linch would answer the same. In UK, for example I remember Open Data advocates ask at most for publication of street names where certain crimes are committed (and they meant crimes happening OUTSIDE homes, like street robbery at gun point, or car theft), never for the names of victims.

      1. Augusto Herrmann

        I agree that there’s much value in opening up some anonymous crime related data, such as the ones you just mentioned in your comment.

        However, I argue that that specific use case of open data (preventing a repeating aggressor from later killing his victim) is not an open data case at all, because prevention is not possible without identifying both the victim and the perpetrator. So it’s more of a data interchange or data mining issue. Not that it’s any less important.

  3. Nitai

    Hi Marco,

    I agree with you. There is lots of ways to analyze anonymous data. To the general public it’s useful to know that “woman attacked would be murdered” is a proved statement. Even better is to know that it happens in the neighbor. Of course for the public security agencies to implement improvements on security policies, in this specific case, they need to know names and address. The next time such woman is attacked the authority may act to avoid a murder.

    Two weeks ago, under a contest named Desarrollando America”, the winner team, who happens to be brazilian, developed an web application using security data from Rio Grande Do Sul also. The name is “where it happens” http://www.ondeacontece.com.br . It shows a map and some numbers about crimes by year and by city.

    Cheers,
    Nitai Silva

    1. marco Post author

      Hi Nitai!

      To the general public it’s useful to know that “woman attacked would be murdered” is a proved statement

      that’s exactly (half of) my point. It is useful to give proof of such statements (without giving names of victims, of course) to the general public to justify, for example, the budget for hiring more counsellors.

      But it is also very useful, I believe, to those counsellors when they try to convince a first time victim that she needs their assistance. Denial can be a bigger obstacle than fear.

      The exception to the “no names” rule may be (but I’m not sure of this, I only provide this as food for thought) already murdered people and already known killers. If TV and newspapers have already said to the whole country that Mrs so-and-so was beat to death by Mr X, for example, collecting in an official website all the names and pictures of such victims with the number of times and period they had suffered violence from who would have become their killer may have a great value for prevention/education.

      What do you think?

      1. Nitai

        Hi, Marco!

        First let me say sorry for the late answer. Augusto and Christian are on vacation, so I’ve been very busy.

        I definitely agree with you. It’s a reasonable way to justify more budget.

        But I’m not sure where is the threshold. Does not have the same impact if they can say Mr X beat his wife to death? Perhaps if we showed their names will inhibit them from seeking assistance. I believe that it’s a sensible situation, and the woman have the trend to keep it secret.

        This discussion will come up more each time we talk about open data.

        Cheers,
        Nitai

        1. marco Post author

          I’m not sure where is the threshold. Does not have the same impact if they can say…

          that surely depends on local culture and customs. As an example (of how much the reaction of women and men from different countries to gender-related “violence” changes) I suggest reading this other article of mine and the comments to it (and of course if you and/or other men and women from Brazil added your point of view it would be wonderful!!!): Humiliating one’s ex-partner online: are men the only ones to do it?

          What I am trying to say is that if we both find the general principle valid, we shouldn’t be surprised anyway if we diverge on how practically implement it, because that can vary a lot from country to country. If I were given the task of writing practical, detailed guidelines on this for any other country than Italy, I would be much less sure, simply because my knowledge of the local culture would range from null to partial

          1. Nitai

            Hi Marco,

            I’m not surprised about our divergence.
            You probably have more accuracy to write those practical and detailed guidelines on how to implement it for Italy than any other person from different country.
            And you’re right again, culture always matters. If the goal is to create any kind of global rules and guidelines to help find thresholds between the very useful information for the general public and the individual privacy rights, it won’t be easy at all.
            I really agree with you that there is contless benefits in using data to help government to save more lifes. Beyond local culture we need to consider local laws.

            Anyway, in our project INDA we avoid get in discussion like this. We have been strictly clear about our scope, we are building “technologies” to help data to be opened.

      2. Augusto Herrmann

        Hi, Marco!

        Chewing on your “food for thought”, even that case (already murdered people) would be problematic, notwithstanding cultural issues, because attention would have to be paid for due process. The fact that someone’s been accused of murder does not necessarily mean that it did happen. What about the cases of false accusations? Indeed, the press sometimes does publicize cases of false accusations as if they were true, this ruining people’s reputations and lives. But they do so at their own risk, and may have to face the consequences.

        So, in our thought experiment, an open government initiative that did this would have to wait until someone is convicted by a court before they would release the data. While there’s some value in that, it would fail one of the 8 principles of open data, that it should be timely.

        So I leave you and your readers with some more food for thought. The threshold where open data meets privacy safeguards is always a thorny issue.

        1. Nitai

          Hi Guys,

          Some more food for thought…

          I was browsing looking for data to catalog in our portal an than I found this http://bit.ly/v5Da5I
          There is lots of PDF’s within the names of murdered people in my home state, Pernambuco. The site is the oficial portal of State Secretariat for Social Defense of Pernambuco.

          I have nothing more to say, just that we have much more to learn.

  4. Nitai

    Hi Marco,

    I’m not surprised about our divergence.
    You probably have more accuracy to write those practical and detailed guidelines on how to implement it for Italy than any other person from different country.
    And you’re right again, culture always matters. If the goal is to create any kind of global rules and guidelines to help find thresholds between the very useful information for the general public and the individual privacy rights, it won’t be easy at all.
    I really agree with you that there is contless benefits in using data to help government to save more lifes. Beyond local culture we need to consider local laws.

    Anyway, in our project INDA we avoid get in discussion like this. We have been strictly clear about our scope, we are building “technologies” to help data to be opened.

  5. Pingback: Open Data for crime prevention: some food for thought from Brazil … - Jornal do Brasil

  6. Pingback: Open Data for crime prevention: some food for thought from Brazil …

  7. Pingback: Open Data for crime prevention: some food for thought from Brazil … : Revista Veja Brasil

  8. Pingback: Open Data for crime prevention: some food for thought from Brazil … | Veja, Brasil

Comments are closed.