blahaibrahim/algerian-seismic-risk-analysis

# Algerian Seismic Risk Engine ### A Geospatial Intelligence System for identifying seismic "Blind Spots" in Northern Algeria. ## 1. Overview The **Algerian Seismic Risk Engine** is a data mining and geospatial analysis pipeline designed to quantify seismic risk in Northern Algeria (Tell Atlas). Unlike traditional maps that focus solely on earthquake magnitude, this system calculates a **Composite Risk Index (CRI)** by convolving seismic energy release with human population density. By mapping data to a high-resolution **H3 Hexagonal Grid**, the system identifies "Blind Spots"—areas with moderate seismicity but high vulnerability (e.g., the Bougara/Mitidja sector) that are often overshadowed by massive but offshore events like the 2003 Boumerdès earthquake. ## 2. Key Features * **Composite Risk Index (CRI):** A normalized metric combining Gutenberg-Richter energy release and WorldPop density data. * **Interactive 3D Maps:** A Streamlit dashboard using PyDeck to visualize risk zones, fault lines, and population centers in 3D. * **Machine Learning Integration:** * **DBSCAN Clustering:** Automatically identifies tectonic domains and active fault zones without prior geological labelling. * **Isolation Forest:** Detects spatiotemporal anomalies (e.g., deep intraplate events or potential precursors). * **Real-time Statistics:** Dynamic calculation of the *b-value* (stress accumulation rate) based on filtered time/space windows. * **Robust ETL Pipeline:** Handles magnitude homogenization (Mw conversion), spatiotemporal deduplication, and reverse geocoding. ## 3. Repository Structure ```text Algerian Seismic Risk Engine/ ├── data/ │ ├── raw/ # Raw ISC CSVs and WorldPop │ └── processed/ # Cleaned datasets and Risk Maps ├── src/ │ ├── etl.py # Pipeline for cleaning, H3 indexing, and CRI calculation │ ├── calculations.py # Seismological formulas (G-R Law, Energy) │ └── ml_models.py # DBSCAN and Isolation Forest …