Research & models / 2026
Shamba Signal.
A Kenya county-year evidence explorer that makes a completed weather-feature experiment understandable.
01 / Context
The problem behind the work.
A plausible model is not automatically a better model. The useful result is the comparison people can inspect.
02 / Contribution
What I built.
- 01
A responsive evidence surface for county-year maize-yield context and experiment results.
- 02
A direct comparison of a weather-feature Ridge model with a county historical-mean benchmark across 47 counties.
03 / Where it stands
The weather model scored 0.337 t/ha versus 0.300 t/ha for the historical-mean benchmark, so the feature did not win. This is a closed experiment, not an operational forecast.
The work in view
Closer inspection.
Open any image for the original capture.

Technology and methods
PythonRidge regressionEvaluationResponsive web
