The scenario
A GIS team needs a reproducible pipeline to predict soil erosion risk from satellite imagery and rainfall data using machine learning, with model validation and map generation.
What is in this drawing
Read the decisions behind it.
01
Choosing whether to impute missing rainfall before feature engineering
02
Deciding when model accuracy is sufficient for production mapping
When building ML pipelines for environmental risk mapping with satellite and climate tabular data.
More like this