Monitoring
Carlos Mendez
Graduate School of International Development
Nagoya University
1
Location: Rio de Janeiro, Brazil. Source: @gustavomellossa via Canva
2
Location: Sao Paulo, Brazil. Source: Tuca Vieira via El Pais
3
5
6
7
9
Economic
geography (EG)
Data
science
(DS)
Geospatial information
science
(GIS)
10
1
Joint work with
Ate Poortinga
Spatial Informatics Group
Nyein Soe Thwa
Spatial Informatics Group
Andrea McMahon
Spatial Informatics Group
Theara Khoun
UNDP Cambodia
Ivan Gonzalez
UNDP Cambodia
1
Geospatial
data
Local
survey
Our
approach
Machine
learning
Cambodia socieconomic survey is useful, but ...
Every 2 years
Sample: 10,000 households
Spatial distribution of the survey
... is costly and has space-time coverage limitations
Cambodia socieconomic survey + BIG earth observation data ...
... less expensive and enhanced coverage
Earth observation data can help us study poverty ...
Provinces
Districts
Communes
... at multiple geographical scales or administrative levels
Earth observation data (satellite and GIS layers) can...
Spatial distribution
of survey data
Enhanced results based on
EO data and machine learning
... enhance our understanding of multidimensional poverty
Survey data + Daytime satellite images + GIS layers ...
Nighttime lights map
Multidimensional poverty map
... can greatly enhance what we know from nighttime lights
Economic
geography (EG)
Data
science
(DS)
Geospatial information
science
(GIS)
Carlos Mendez
Graduate School of International Development
Nagoya University