Mafuriko is the Swahili word for flood. It is also the name of the first tool in the resilience stack — a flood intelligence system built for urban Kenya, where the relationship between rainfall, drainage infrastructure, and human settlement has become increasingly volatile.
Nairobi floods predictably every long rainy season. The Nairobi River system, the seasonal streams running through informal settlements, the drainage channels designed for a city of one million now serving five — these structures fail. What is less predictable is which neighbourhoods flood first, how fast the water moves, and whether the people downstream have any warning.
What Mafuriko Does
Mafuriko aggregates rainfall data from the Kenya Meteorological Department, combines it with river level readings from available gauging stations, layers in satellite-derived topographic data, and produces a risk map at the sub-location level. It answers the question: given current conditions, which areas should be on alert?
The gap in Kenyan flood management is not data availability — KMD publishes rainfall data, the Water Resources Authority maintains gauging stations, satellite imagery exists. The gap is aggregation. Nobody was pulling these sources into a single operational picture at a resolution useful to the community health worker or the Red Cross coordinator. Mafuriko is that aggregation layer.
The tool runs on a lightweight Python backend, outputs to a simple web interface and SMS alert format, and requires no proprietary data sources. All inputs are public. The 2023 and 2024 floods in Mathare are the use case it was built for.
Responses