Slum: Comparing municipal and census basemaps
Slum identification and mapping are fundamental to support public policies. However, slum data is affected by differences in definition, identification, spatial delimitation, sample and data collection periods. This study compares slum maps in Sao Mateus (a peripheral region of Sao Paulo city) from the 2010 demographic census and the Sao Paulo municipal government. The census and the municipal basemaps and orthophotos were overlaid in a Geographic Information System (GIS) to evaluate their similarities and differences, the reasons for inconsistencies between them. Of the 68 districts with subnormal agglomerates (EDSAs; special enumeration districts for slum areas in the Brazilian demographic census), 59% are in the categories ‘high coincidence’ or ‘partial coincidence’ with the perimeters of favelas (the main type of slum in Brazil) delineated by the Sao Paulo municipal government; the names of the residential area (favela/EDSA) are the same in only 51.5% of the subnormal agglomerates, while 26.5% of the EDSAs include residential areas not considered favelas by the municipality. Many Brazilian researchers state that the underestimation of EDSA census data is accentuated by the exclusion of favelas smaller than 51 dwellings from the census. However, at least in Sao Mateus, the exclusion of small favelas (3.6%) was not the only reason for census underestimation of favelas.