This model aims to derive demographic and socioeconomic characteristics using a set of covariates from VHRRS and HRRS for improving our understanding of the spatial demographic and socioeconomic patterns in SSA cities and their implication for malaria transmission risk. As far as population density is concerned, at VHRRS the results obtained in MAUPP on Dakar and Ouagadougou will be used and the methods developed will be applied to the two other cities: geo-statistics will be used to study the relationships between population data derived from censuses or surveys and LU/LC classes. Based on these relationships, population weights will be assigned to the different classes and populations will be redistributed accordingly. However, we will go beyond MAUPP as we will use population density and building volume to derive a crowding index. Besides, geo-statistical methods will also be applied to explore the relationships between remote sensing variables and a synthetic socioeconomic index calculated based on different socioeconomic indicators from survey or census data. The output will serve as input to the epidemiological risk modelling.

The intra-urban variation of crowding and socioeconomic status will be estimated in a 100m grid based on 1) survey data available for clusters (and/or census data if possible) and 2) VHRRS data. Moreover, the potential of free HR Sentinel-2 data also will be explored for deriving the crowding and socioeconomic status indices.

Crowding index from VHRRS

Population density will be estimated using the methodology developed in MAUPP (for the 2 cities where it is not yet available): Statistics will be used to study the relationships between population data derived from censuses or surveys and LU/LC classes. Based on these relationships, population weights will be assigned to the different classes and populations will be redistributed accordingly.

  • Building floor area based on building surface, height and storey number will be estimated based on 3D products derived from Pléiades tri-stereo imagery.
  •  Population density per building floor area will be calculated to simulate crowding.

Crowding index from HRRS

  • A spectral shadow index will be computed on VHR and HR images.
  • Statistical relationships between building floor area, VHR shadow index and HR shadow index will be explored.
  • If the statistical relation between the building floor area and the HRRS shadow index is significant, population density per building floor area will be calculated to simulate crowding from HRRS.

Socioeconomic status index from VHRRS

In developing countries, there is a strong urban socioeconomic segregation in space, which means that cities are structured in neighbourhoods where households have similar socioeconomic characteristics. Besides, the morphology of neighbourhoods, as interpreted from remote sensing, can reflect these characteristics.

  • We will carry out multi-variate statistical analysis to investigate the relationships between (i) variables derived from VHRRS in WP3 (average building size and height, spectral and texture indices, percentages of LU/LC classes, type and importance of road network, presence of built-up on/close to unsuitable areas (steep slopes, lowland flood plains, marshy areas, dumpsites…), distance to city centre (i.e. to employment opportunities), etc.) and (ii) socioeconomic variables from surveys (and/or census of available).
  • Based on the results, we will produce a synthetic socioeconomic index.

 Socioeconomic status index from HRRS

  • We will explore the potential of HRRS Sentinel-2 data for carrying out the same type of analysis as in 8.3 and producing a synthetic socioeconomic index.