Quantifying gang activity in an urban environment

‘Guns of Brixton: which London neighborhoods host gang activity?’  by Alessandro Venerandi, Giovanni Quattrone, and Licia Capra
Best Paper Award at Urb-IoT 2016, 2nd EAI International Conference on IoT in Urban Space

Street crime and gang activity is a complicated subject that is related to a wide range of economic and social issues, only one of which is the level of education. It has proven to be very difficult to reach widely applicable conclusions as these conditions may vary wildly from neighborhood to neighborhood, not to mention from country to country. But there are ways of framing this subject within rigid, quantifiable data. Authors of this paper have successfully quantified the relationship between urban form (density and connectivity of built environment, and house typologies) and crime activity by using Open data as input and by taking the metropolitan area of London as case study.

As the they elaborate: “We identify a set of descriptors which are able to capture multiple aspects of the built environment such as build period, type and fiscal band of dwellings, place accessibility, and street network connectivity. We extract measures to describe the built environment from Ordnance Survey (OS) VectorMap District and the London Datastore (i.e., the official web portal for statistical information on Greater London). We derive the areas where gangs currently operate from a geodataset made publicly available on an online article by the Independent. The information contained in that dataset was gathered through an interview which saw the involvement of an ex-sergeant of the London Metropolitan Police. Having computed the measures and assigned presence or absence of gang activity to the different areas of London, we perform step-wise logistic regression to obtain a parsimonious set of descriptors of the urban environment which are significantly associated with the activity of criminal groups.”

This brings us closer to understanding how crime develops in neighborhoods, indirectly pointing us towards ways of lowering compatibility with crime by design. If you would like to learn more, you can find the full paper for free here.