Yeto4president/SahelSafe

SahelSafe is a machine learning project aimed at predicting high-risk zones and periods for terrorist attacks in Burkina Faso, using open ACLED data. # SahelSafe ## Introduction **SahelSafe** is a machine learning project aimed at predicting high-risk zones and periods for terrorist attacks in Burkina Faso, using open ACLED data. Leveraging an XGBoost model, interactive visualizations, and a "Data for Good" approach, this project seeks to provide alerts and risk maps to support security and humanitarian efforts in the Sahel region. Last updated: May 23, 2025, 11:20 PM CEST. ### A Mission Born from Personal Commitment I was born in Burkina Faso. It's always been a peaceful country. Until the day when, in total incomprehension, terrorists started attacking certain parts of the country and even once the capital. For a while it didn't affect me directly, but some of my relatives were affected by the death of several of their acquaintances in the attacks. This crisis has several consequences: thousands of people are fleeing danger, the economy is stalled by instability and an impatient population is attacking its leaders. Above all, thousands of families are shattered by the loss of their loved ones. When I began exploring conflict data in Burkina Faso, I didn’t just see numbers and coordinates—I saw *lives impacted*, communities uprooted, and an urgent need for solutions. Terrorism, which has shaken my country in recent years, is more than a security challenge; it’s a human crisis that demands action. This predictive analysis project was born from that realization. I wanted to use my data science skills not for an academic exercise, but to contribute, however modestly, to *changing things*. ### A "Data for Good" Approach Armed with ACLED’s open data and a passion for societal impact, I built a machine learning model to forecast high-risk zones and periods for attacks. It’s not just code or a map—it’s an attempt to give decision-makers, NGOs, or local communities tools to anticipate, prepare, and perhaps *save lives*. By integrating interactive visualizations and probability-based alerts, I aimed to turn ra …