Aral System: A Monitoring Platform for the Aral Sea Region
Our group’s own product: the platform aggregates satellite and climate data for the Aral Sea region, forecasts dust storms with ML and sends out alerts. Seven languages, open access.
The challenge
Where the Aral Sea used to be lies the Aralkum desert, and its dust storms hit the health of people in Karakalpakstan and neighboring areas.
Data about the region exists — imagery, climate models, reanalyses — but it is scattered across scientific sources. We brought it together into one public platform.
What we built
- An up-to-date dashboard of air quality and weather conditions across the region.
- Satellite monitoring of the Aral Sea on Landsat imagery: how its area and condition change.
- Integration of scientific sources: Open-Meteo (weather data), GloFAS (hydrology), CMIP6 (climate scenarios).
- ML forecasting of dust storms and dust levels.
- Telegram alerts when air quality deteriorates.
- Interface and content in seven languages — the platform is accessible to audiences across the entire region.
Engineering challenges
- Heterogeneous sources: datasets in different formats and at different frequencies — we built a normalizing pipeline.
- Heavy geodata: satellite layers are preprocessed ahead of time, so the dashboard stays fast.
- Seven locales designed into the architecture from day one, not bolted on at the end.
The team's role
Data pipelines, ML, backend, frontend, alerting, localization — all built in-house.
It is a product, not a client order: the best answer to the question of how we work with complex data and a new domain.
Need a platform built on complex data?
We'll show you how we'd approach your challenge: break down the data sources, propose an architecture and a plan. The first consultation is free.