Modelling urban deprivation in Kinshasa, DRC

We modelled and predicted an urban deprivation score for the 23 communes of Kinshasa province and stored in urban_deprivation.tif. It is the outcome of a confirmatory factor analysis stored in code_simplified.R that combined four geospatial covariates stored in the covariates folder: distance to river and residential road density in a 1km window derived from OpenStreetMap and building landscape shape and building area coefficient of variation derived from a building footprint layer. The model structure was validated through several metrics that indicate a good reproduction of the correlation matrix between the covariates at study sites. The predicted gridded score was then compared with qualitative information collected from the litterature. We provide for comparison this mapped information in qualitative_info.shp where the source is stored in the attribute table and matches the bibliography stored in qualitative_info_source.txt.