The State of Brazilian Favelas
Description
This dataset provides a national-scale, harmonized database of all Brazilian favelas ("Favelas e Comunidades Urbanas", FCU) delineated by the 2022 Brazilian Census (IBGE), covering 12,344 informal settlements. It integrates census-based socio-demographic and housing indicators with environmental, topographic, air-pollution, and accessibility variables derived from multiple geospatial sources. Main integrated files. The primary analysis-ready layer is shp/final_version.shp (one row per favela, ~89 fields), with shp/final_version_no_missing_values.shp as the complete-case subset and json/final_version.json a GeoJSON export. The clusters/ layer adds the infrastructural-regime labels. Attributes follow a multi-scalar framework: Identifiers & geography: favela, municipality, and state codes/names (CD_FCU, NM_MUN, SIGLA_UF, ...), area, population, and density. Household level (socio-demographic & housing): household density, ageing index, median age, sex ratio, literacy, race/colour composition, socio-economic deprivation (CIDACS Brazilian Deprivation Index, IBP), occupation and dwelling types, and detailed sanitation, bathroom, and water-supply/source categories. Within-area level (environmental & hazard): water and vegetation cover, risk areas, slope classes, Height Above Nearest Drainage (HAND) classes, air-pollutant concentrations (SO2, CO, O3, NO2; Sentinel-5P), waste-collection coverage, and establishment/economic-activity densities by type. Area-connect level (accessibility): distances to the nearest primary-care unit and inpatient hospital. Composite indices and regimes. Four indices aligned with UN-Habitat definitions are provided - improved water access (wtr_idx), improved sanitation access (sanit_idx), sufficient living area (liv_a_idx), and housing durability (hh_idx) - together with a cluster label (cluster) identifying three infrastructural regimes derived by unsupervised K-means clustering: Structured (0), Partially Structured (1), and Unstructured (2). Supporting source and intermediate layers. The deposit also includes the raw and harmonized inputs: 2022 and 2010 IBGE census sectors, the CIDACS deprivation index (IBP), MapBiomas 2023 land cover, per-favela Sentinel-5P air-pollution results (one file per settlement), AGSN 2019 health-facility layers and computed hospital distances, IBGE 2022 municipality and 2010 locality boundaries, and the raw household census tables. Coordinate reference system. All layers are referenced to the official IBGE favela polygons and harmonized in SIRGAS 2000 (EPSG:4674); the integrated final_version export is provided in Web Mercator (EPSG:3857). This dataset supports research on urban inequality, informal settlements, infrastructure provision, environmental exposure, spatial analysis, and machine learning. Associated codebase: https://github.com/dev-jotape/state_of_brazilian_favelas.