Arvato Systems' Crime Map Goes Public
From a Proof of Concept (POC) to a Real Product
Arvato Systems' Crime Map Goes Public
In our January 2019 newsletter, we introduced our “AI text classification of crime news” POC helping to analyze and visualize big data. We are very proud to announce that our Crime Map POC developed into a real product used to analyze and visualize crimes of one city in Germany.
About the Project:
Burglary, theft, accidents, or even violence is unfortunate but very common in big cities. That is also the case for the sixth most populated city in Germany, Stuttgart. Instead of ignoring the problem, the “Stuttgarter Zeitung” and “Stuttgarter Nachrichten” (news outlets) started an initiative last year and launched the Crime Map this summer to create more transparency. Thus, the news outlets are using our Crime Map to analyze and visualize the crimes announced by police reports published by the Police Headquarters in Stuttgart.
The Crime Map is an example of exclusive data journalism that uses Artificial Intelligence to create extra content and new approaches. Also, the Crime Map Solution creates more transparency for the private end-user. Users, e.g., tourists looking for hotels or residents who are looking for a new apartment or home can use the database to search areas for occurred crimes. Not only the crime itself can be looked up, but also the frequency crimes happen in a particular area. Even though the project was only launched recently, data reaching back to 2014 can be accessed and visualized with the help of the map, showing the development of crime in particular areas over time.
From the beginning, the Crime Map was supported by an initiative of the “Südwestdeutsche Medienholding” named “Mission Innovation.” Through the program, the project partners were brought together, and the project was financially supported. In the official Press Releases (German only) the Corporate Entrepreneur at SWMH Tobias Köhler stated: “The Crime Map shows that sometimes it only takes a small push to make great projects fly.”
How does the solution work?
By using Artificial Intelligence (AI), the algorithm behind the Crime Map automatically classifies the full-text messages (usually consisting of approx. 12 sentences) into several relevant categories (such as theft, accident, traffic, violence, etc.) and transfers them to a database. Note that different from other states, German police offices do not provide their crime statistics to the public directly. The information is only available via full-text press releases; hence, the necessity to implement an AI algorithm to extract the relevant crime categories of these texts. Thus, our Crime Map is automated to recognize and then list out crimes in real-time onto a city map. Even though the project was only launched recently, data from 2014 onward can be accessed and visualized with the help of the map.
If you are interested to see the real application of our solution please visit https://www.stuttgarter-zeitung.de/crimemap (unfortunately the website is only available in German).