Sign In

Communications of the ACM

ACM TechNews

Uncovering Crime Patterns Using Location Data

View as: Print Mobile App Share:
Satellite image of Philadelphia.

Where the researchers found more mobility based on the location data, they also found more crime.

Credit: Halina/

Researchers at ETH Zurich in Switzerland, the U.K.'s University of Cambridge, and New York University have demonstrated a method for inferring crime patterns, using anonymized data from location technology platforms.

The team used data from the Foursquare platform to calculate mobility flows in San Francisco, Chicago, and Philadelphia from 2012 and 2013.

The researchers compared these datasets with statistics on criminal offenses relating to theft, robbery, assault, burglary, and stolen vehicles; the more activity the data indicated for a specific time and district, the greater the number of offenses committed there.

Activities included actively shared user locations (check-ins), with paths between check-ins calculated on the basis that users choose the shortest route and navigate using existing traffic routes.

By training different prediction models with Foursquare datasets and with data on past offenses, the researchers learned that those fed mobility data could better predict crimes than those trained solely on previous crimes.

From ETH Zurich (Switzerland)
View Full Article


Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account