Satellites and AI for water and mining: hands-on workshops

IC-TWEP 2026 · University of Tlemcen · 26–29 October 2026 · presented by Dr Firas Alazem · materials by QuestFeed. Two one-hour workshops with Google Colab notebooks you can keep using afterwards.

Open a notebook in Google Colab. Download the notebook, go to colab.research.google.com, choose File → Upload notebook, then Runtime → Run all. A Google account is all you need: the notebooks read open satellite archives and the QuestFeed Open Data Store directly.
Workshop 1

Watching reservoirs from space

Measure reservoirs with open satellite data: today's water surface, ten years of monthly change, rainfall and drought upstream, water-quality signals and an AI change radar.

Watching reservoirs from space: satellite view used in the workshop
Live workshop notebook

Reservoirs from space

Choose a reservoir, map today's water live from Sentinel-2, follow ten years of change, link it to rainfall and drought, read water-quality signals and run the AI change radar.

Run time in Colab: about 2 minutes
Take-home

Erosion and siltation risk

Where a reservoir's sediment comes from: the RUSLE soil-loss model from open rainfall, soil, terrain and land-cover data, with an order-of-magnitude siltation estimate.

Run time in Colab: about 8 minutes
Take-home

Floods through the clouds

Map the 2022 floods at the Niger–Benue confluence with Sentinel-1 radar, which sees through clouds, then try a flood you know.

Run time in Colab: about 3 minutes
Take-home

Your own reservoir monitor

Any reservoir on Earth: a monthly open-water record and report files, computed with the workshop's method.

Run time in Colab: about 5–10 minutes for 12 months
Workshop 2

Minerals, mining and water

Read the mineral make-up of rock from orbit, test what each sensor can see, use an AI foundation model, and follow mines and rivers through time.

Minerals, mining and water: satellite view used in the workshop
Live workshop notebook

Minerals, mining and water

Mineral fingerprints, one district seen by four sensors, AI similarity search and mapping with an honest test, mines growing since 1990, and mining's mark on rivers.

Run time in Colab: about 3 minutes
Take-home

Transparent prospectivity mapping

A knowledge-driven prospectivity map for an Algerian gold district, with every weight visible and an honest check against known mining areas.

Run time in Colab: about 4 minutes
Take-home

Hyperspectral deep dive

Work with real EnMAP spectra: absorption depths, spectral angle mapping and a comparison with NASA's EMIT mineral identifications.

Run time in Colab: about 1 minute
Take-home

Structures from elevation data

Lineaments, their directions and density from the Copernicus elevation model, and how alteration gathers around them.

Run time in Colab: about 1 minute