The project
The MAUPP project (Modeling and forecasting African Urban Population Patterns) is a four-year research project (2014-2018) funded by the STEREO-III program of the Belgian Science Policy (BELSPO).
The main objective of MAUPP is to develop models and forecasts to improve our knowledge of sub-Saharan African urban population patterns for a wider usage of urban population distribution datasets in vulnerability and health assessments. The MAUPP project aims to contribute to the WorldPop open access archive of spatial demographic datasets.
Built-up and population maps at moderate spatial resolution (12.5m) will be produced for 48 african cities and for five distinct years: 1995, 2000, 2005, 2010 and 2015. Very high spatial resolution built-up and population maps will also be produced for 3 cities: Ouagadougou (Burkina Faso) and Dakar (Senegal).
Project website: Click here.
My achievments
In this project, I conducted my PhD research which leads to the development of a complete semi-automated framework for mapping land cover, land use and model human population density. All these developments were made using open-source GIS software and Python.
Skills learned and/or deployed:
- Optical VHRRS image processing, object-based image analysis (OBIA), including cutting-edge methodological developments for scaling of traditional segmentation approaches
- Python programming (incl. Pandas, Numpy, Scikit-learn)
- Machine learning (Random Forest, Xgboost, hyperparameter optimization, feature selection)
- Automation of Geospatial processes
- Big data management and processing
- Postgres/PostGIS, GRASS GIS, QGIS, R
- Webmapping (Leaflet)
- HTML/CSS (GoHugo static website)
- OpenStreetMap
- Scientific publishing
- Open science
- Project management