Work

Predicting Crime in London: Comparing Streetscape Environmental Features Extracted from Images and Social Features

Year: 2025

Type: conference-paper

Abstract: Crime rates are influenced by social and streetscape environmental features. Although advancements in machine learning have made it easier to measure these features, many studies have overlooked their predictive potential. To address this gap, this study first randomly sampled 1,000 LSOAs in London, and extracted data on these features to explain variations in crime rates. Then, various models were used to predict crime rates at an additional 500 LSOAs. The findings suggest that allowing for geographical variations is important. Social features to be the most important predictors of crime, but for some crimes, including streetscape environmental features can enhance model’s accuracy.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.15230108

Authors: Guo, Sitong, Harris, Rich, Zhu, Rui

Institution: University of Bristol

Language: English

Topic: Crime Patterns and Interventions

Subfield: Sociology and Political Science

Field: Social Sciences

Domain: Social Sciences

SDGs: Sustainable cities and communities, Peace, justice, and strong institutions

Open Access status: green

OpenAlex ID: W6949376025

FWCI:

Cites:

Cited by:

Related to: