As part of the WDAqua R&D Week in Berlin, students were asked to participate in a hackathon which used open data to generate interesting insights.

In this ‘pitch’, Hady El-Sahar, Laura Koesten, Pavlos Vougiouklis and Denis Lukovnikov talk about their project iCrime, which playfully explores how open data could be used to make crime easier to commit!

We’ve all heard stories of people getting caught for committing a particular crime. But getting yourself locked up behind the unknown and potentially unfriendly environment of a prison cell just because you fancy yourself in the crime business can be quite demotivating. We are here to successfully free you from the cold-hearted fear of ending up behind the bars. At this point you are probably wondering how? And until the R&D Week I would have probably shared the same view.

Truth is that with the rise of the open data initiatives, all the data about arrests, prosecution, and random police stop-and-search along with the demographics of the LSOAs is freely available to those that would be keen to experience the thrills of committing a crime. Our idea consists of triangulating the general population statistics along with crime-related data to not only suggest the most appropriate crime but also help you escape (or not?) to your route to freedom. Missing some fellow criminals in prison? We can arrange you a short visit there.

Despite what we know from the media, the bias when it comes to stopping or arresting people due to their ethnicity was still slightly shocking. While we ended up presenting iCrime as a “plan your perfect crime planner”, we developed the prototype with other use cases in mind. For instance using what we know about crime statistics as a bias awareness tool which can be used by police officers or the use of the “escape route planner” as a way of suggesting an especially safe route through the city.

“Let’s aim for the funniest prize”, was our strategy for winning something during the R&D Week. Since we dropped the glorious attention-is-all-you-need sequence-to-sequence models with COPY actions, our chances of winning the technical prize is basically non-existent.

Making a funny story out of criminal records dataset can be quite challenging. Some team members suggested dropping this angle and making an app to tell us how many WWW papers we’d publish so far. But in the end we stuck to the plan by creating the iCrime app that helps criminals to commit a perfect crime and to get away with it. Oh well!

We were able to have fun while working on this data and aimed at presenting a serious topic in an engaging way - which paid off when Jeni Tennison (CEO of the Open Data Institute) presented us with the prize for the most ‘fun’ project: two cute little fluffy unicorns! A great prize - next step is figuring out how to split this between the four of us!