Kenya Soil Health Data Warehouse
An end-to-end BI pipeline on Kenyan agronomic trial data — data warehousing, OLAP querying, interactive visualisation, and data mining
Project Overview
An end-to-end business intelligence pipeline built on the Kenya Soil Health dataset — on-farm agricultural trials across multiple agro-ecological zones, with grain yield as the target variable. Done for the Business Intelligence & Analytics course at the University of Nairobi.
Key Achievements
- Modeled the warehouse as a galaxy (fact constellation) schema — two fact tables for crop yield and farm inputs sharing conformed location and time dimensions, alongside star and snowflake implementations
- Implemented OLAP operations (slice, dice, pivot, and drill-down/up/across/through) over the warehouse
- Built an interactive Streamlit dashboard with heatmaps, fever charts, and gauges for exploring yields across zones
- Ran data-mining workflows in Orange3 — clustering, pattern discovery, and time-series analysis