The project

This small disseminaiton and support project aimed to produce weighted layers for dasymetric mapping that can be used for spatial redistribution of a response variable into a raster grid with a finer spatial resolution. For the ease of user, we developed with Charlotte Flasse a new GRASS-GIS add-on, called “r.area.createweight”. This simple and convenient ready-to-use tool facilitates the implementation of a Random Forest machine-learning based approach, similar to the one used in the WorldPop project. The tool can be used for a wide variety of spatial phenomena, from the mapping of human or animal populations, to the mapping of other socio-economic or environmental variables.

For more information, please check the ISPRS publication: Click here.

Skills learned and/or deployed:

• Machine learning in Scikit-learn (feature selection, hyperparameter optimisation)
• Python programming
• GRASS GIS
• Git/Github
• Flake8