Fig. 2
From: Advancing data science research education in Africa through datathon-driven innovations
Source and service layer credits for satellite imagery: Esri, DigitalGlobe, GeoEye, Earthstar, Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo, and the GIS User Community.
Pedagogical training examples comparing traditional and AI/ML analog approaches. AI artificial intelligence, ML Machine learning, RMSE root mean square error. Panel (a): Stepwise demonstration using a relational data set. Step 1 introduced the problem with a single data set. Step 2 applied ordinary linear regression, and Step 3 transitioned to an AI/ML learning workflow by splitting the data for training and evaluation of the regression model. Panel (b): Application to satellite imagery. Step 1 presented the source satellite image. Step 2 employed manual digitization (considered as a traditional statistics approach), and Step 3 demonstrated the automated segmentation of housetops using AI/ML techniques.
