Police-Reported Homicides in Barranquilla, Colombia, 2008-2023

This dataset is part of the ESCALA (Study of Urban Health and Climate Change in Informal Settlements in Latin America) project that was funded by the Lacuna Fund of the Meridian Institute https://lacunafund.org/.

This dataset contains registered victims of homicides in a neighborhood within the municipal area of Barranquilla, 2008-2023. The data were provided by the National Police from the SIEDCO database (Sistema de Información Estadística, Delincuencial, Contravencional y Operativo). The original dataset had the following potential errors: (1) anomalous values in the column "quantity", which was expected to be 1 as each record represents a victim (0.9% of all records), (2) 41 records with incomplete data, i.e. only include the victim's age but no other data about the victim or the agressor, (3) some of the reported neighborhoods are not in the official neighborhood layer.

Data cleaning included: (1) records with anomalous values in the column "quantity" were kept and assumed to represent a victim each, (2) records with no information other than the victim´s age were assigned NA for other attributes, (3) records with neighborhoods not found in the official neighborhood geographic layer were checked against secondary sources and assigned to the corresponding official neighborhood. Specifically, "Barranquillita" and "Bendicion de Dios" were assigned to "Villanueva"; "Moderno" was assigned to "VillaBlanca";"Conidec" was assigned to "Ciudadela 20 de Julio"; "Boliche" was assigned to "Centro"; "San Martin" was assigned to "Carrizal"; "Puerta Dorada" and "Sevilla Real" were assigned to "Las Granjas"; "Sinai" was assigned to "El Silencio". The original neighborhood was kept in the attribute COMMENTS. Records with the following neighborhoods were assigned NA in the attribute ID_NEI, as those neighborhoods are not in the official neighborhood layer, such as: "Pinar del Rio", "Alameda del Rio", "Villas de la Cordialidad", and "Caribe Verde". Other records with NA in the ID_NEI attribute indicate that there was no neighborhood information in the original dataset.