Inspiration
Dandora is one of Nairobi’s largest dumpsites, and pollution around the area can affect nearby communities. Traditional environmental monitoring often depends on ground-based sensors, which can be expensive to install, maintain, and scale across large areas.
At the same time, satellites are already collecting environmental data for free. The problem is that this data is often difficult for non-specialists to interpret and turn into something actionable.
That inspired Kenya AirWatch: a way to transform free satellite observations into a simple environmental early-warning tool.
What it does
Kenya AirWatch monitors atmospheric methane around Dandora using Sentinel-5P satellite data.
Instead of showing users raw satellite measurements, the system processes historical observations, establishes a baseline, and detects unusually high readings that may deserve further investigation.
The dashboard lets users:
- View methane observations over time
- Identify unusual spikes and anomalies
- See the monitored area on a map
- Quickly understand whether conditions appear normal, elevated, high, or severe
The goal is not to replace ground sensors, but to help environmental teams decide where and when further investigation may be needed.
How we built it
Kenya AirWatch uses Sentinel-5P/TROPOMI satellite observations accessed through Google Earth Engine.
A Go ingestion service processes the observations and stores the cleaned measurements in PostgreSQL. A Go REST API then exposes the data and runs anomaly detection using historical averages, standard deviation, and z-scores.
The frontend is built with React and TypeScript, with Leaflet for mapping and Recharts for visualizing methane trends.
The overall flow is:
Sentinel-5P → Google Earth Engine → Go ingestion → PostgreSQL → anomaly detection → React dashboard.
Challenges we ran into
One of the biggest challenges was making sure the project did not overstate what satellite data can tell us.
Sentinel-5P can show methane concentrations in the atmosphere around Dandora, but it cannot prove that a specific methane reading came directly from the dumpsite. Wind, nearby sources, and atmospheric conditions can all affect measurements.
Another challenge was converting scientific satellite data into something understandable and useful without losing its meaning.
What we learned
We learned that collecting data is only one part of environmental monitoring. The real value comes from transforming that data into information that helps people decide when something deserves attention.
We also learned how to work with Earth-observation data, build anomaly detection around historical baselines, and design a system that communicates scientific uncertainty clearly.
What's next
We plan to combine data from multiple Earth-observation satellites, including NASA and ESA missions, to improve coverage, cross-check observations, and make Kenya AirWatch more reliable.
Future work could include weather data, additional pollutants, multiple monitoring sites across Kenya, automated alerts, and validation with environmental professionals.
Log in or sign up for Devpost to join the conversation.