Kenya Connectivity

Methodology & Data Sources

The Kenya Connectivity Map combines six data layers, four of which roll up into a single composite score. This page explains how that score is built, what each layer measures, where the data comes from, and where it falls short.

Composite Score Formula

Each dimension is normalised to a 0–100 scale before weighting. Internet usage carries the most weight since it is the most direct measure of digital access. Electricity is weighted lower because it is a precondition for connectivity, not a connectivity metric on its own.

Data Sources

Known Limitations

• Tower density is crowd-sourced and may undercount towers in low-population areas. Treat as indicative, not comprehensive.
• Mobile speed reflects users who ran a speed test, which skews toward urban and higher-income users. Rural speeds may be overestimated relative to typical experience.
• Starlink bias: a handful of Starlink users in a county can inflate its average speed. Read speed alongside tower density to separate ground-based from satellite coverage.
• Electricity (2019): county-level figures are still the 2019 census, the most recent public breakdown available. National access has since risen to roughly 76% (KPLC, 2025), but no newer county-level split has been published.
• Per-school internet connectivity is not yet available. Giga's School Profile API covers this; access has been requested and is pending approval.

License

Code: MIT License. Data layers keep their original licenses, KNBS and DigiSchool data is Kenya government open data, OpenCelliD is CC BY-SA 4.0, ENERGYDATA.INFO datasets are CC BY 4.0, Giga school data is ODbL (credited to Giga and OpenStreetMap contributors), and OCHA boundaries are Common Operational Datasets.

Full README & source on GitHub →