About the Project

Latin America and the Caribbean (LAC) remains the most violent region on the planet, accounting for over a third of homicides in the world. Increasingly, data and information are being promoted as a powerful tool to understand and prevent crime and violence. However, there are two key questions that remain largely unanswered:
1) Why is crime clustering in certain neighborhoods?
2) Why do certain individuals turn to crime and not others?

Leveraging traditional (official statistics, surveys, interviews) and new data sources (big data, i.e CDRs and open source GIS), Ciudata Segura aims to build a granular spatio-temporal tool to diagnose crime factors in cities and better inform security policy-making.

Pilot in

The pilot phase is currently being deployed in six Colombian cities presenting a diversity of contexts, both in terms of socioeconomic and political characteristics and crime prevalence: Bogotá, Medellín, Barranquilla, Montería, Valledupar, and Tumaco. Preliminary results will be shared with our local partners by the end of 2018.

A Closer Look

The big data models used in this study have been previously tested in three cities, including Boston, Los Angeles and Chicago. The models include the following variables:

  • Physical characteristics of the city,  such as land use, size of blocks,  age of buildings, population density, vacuums
  • Socioeconomic characteristics,  such as unemployment and income inequality 
  • Mobility patterns of individuals

Overall, results have shown that a comprehensive model including all variable above is the best better predictor of crime, but also that crime dynamics cannot be explained to the same extent in all cities.

Learn more about Ciudata Segura:

Learn more about using Big Data for crime prediction:


For more information about Ciudata Segura,
please write to