Omdena-Tanzania-Kesho-AI
# AI-Powered Early Warning System for Extreme Weather in Tanzania
This repository contains the details and outcomes of the grassroots AI Climathon event hosted by Omdena in partnership with EIT Climate-KIC and Sahara Ventures. The event focused on the development of a precision early warning system for predicting extreme weather conditions in Tanzania.
## Table of Contents
1. Summary
2. Problem Statement
3. Goal
4. Approach
5. Implementation
6. Metrics for Success
7. Future Work
8. License
9. Contact
## Summary
We are building a 24-hour AI-powered precision early warning system designed to predict and provide timely alerts for extreme weather conditions in Tanzania. This system will utilize data from the Tanzania Meteorological Authority (TMA) and historical data to predict floods, heavy rains, and heat waves. It will deliver tailored alerts via SMS to individuals, enhancing preparedness and response to weather-related disasters.
## Problem Statement
**Why are we doing this?**
Unpredictable weather conditions caused by climate change lead to severe adverse effects such as floods, heat waves, and other extreme weather events in Tanzania. These conditions result in significant property damage, displacements, and fatalities. There is a lack of tailored preparedness measures, specific location pinpointing, and timely alerts, leading to inadequate disaster response and low accessibility of vital information.
## Goal
### SMART: Specific, Measurable, Attainable, Relevant, and Time-bound
- **Specific:** Develop a robust AI-powered early warning system to predict extreme weather conditions in Tanzania, specifically floods, heavy rains, and heat waves. The system will leverage data from TMA and historical weather data to provide accurate alerts via SMS.
- **Measurable:**
1. Achieve a prediction accuracy of at least 80% for floods, heavy rains, and heat waves within the first eight weeks of model development.
2. Develop and deploy a user-friendly dashboard visualising w …