Georeferenced Crime Dataset for Bogotá, Colombia (2020–2026): Integrated, Cleaned and Normalized Records

This dataset contains georeferenced and cleaned crime records for Bogota, Colombia, covering the period 2020-2026. It was constructed through a systematic data integration and cleaning pipeline that consolidated heterogeneous official sources from Colombian institutions (Policia Nacional, Fiscalia General de la Nacion, SIEDCO), resolving issues of duplication, inconsistency, and missing geolocation.

The pipeline applied the following processes: normalization of event records by crime type and date, deduplication using record linkage techniques, spatial georeferencing using DANE municipal codes and coordinate validation, and feature engineering including one-hot encoding of categorical crime variables. Spatial outlier detection was performed using Haversine distance metrics.

The resulting dataset supports spatial crime analysis, clustering (K-Means, DBSCAN), and hotspot identification in Bogota. It is intended as a research and educational resource for criminology, urban data science, and public policy analysis.

This dataset is a research output from the project: Analisis espacial de la criminalidad en Colombia: un enfoque metodologico basado en la calidad del dato georeferenciado. Universidad Pedagogica Nacional, 2026.

Also available on Kaggle for broader accessibility.