Projects I have successfully completed

1. Kenya Climate Risk Monitor – County-Level Drought & Flood Early Warning

Problem:
Kenya’s 47 counties, especially ASAL regions, suffer recurring drought and flood crises affecting millions. Most early warning systems lack automated, county-specific risk scores. This project solves the lack of real-time climate intelligence by building a pipeline that ingests daily weather data, detects anomalies against 40+ year historical baselines, and surfaces actionable risk scores for disaster preparedness.

Tools/Stack:
Python, Kestra, Terraform, dlthub, Google Cloud Platform (BigQuery, Cloud Storage), dbt, Looker Studio, Docker

Your actions:

• Built end-to-end ELT pipelines with Kestra orchestration and dlthub for ingestion

• Modeled dimensional tables and fact risk scores in BigQuery using dbt

• Optimized aggregation queries to compute rolling 7-day anomalies across 47 counties

Outcome:
Reduced query runtime from 45s to 12s (73% faster). Enabled daily county-level risk dashboards for drought/flood thresholds. Led to actionable understanding of precipitation and temperature thresholds that precede extreme events in Kenya.

Link:
GitHub Repository