ProcrustesBed/african-conflict-analysis

# african-conflict-analysis A Python tool to pull and analyze ACLED conflict data across African countries. I built this primarily to practice the usage of **pandas**. The program automates the process of identifying key actors and mapping out violence trends. ## What it does * **Fetches an ACLED dataset** automatically via `kagglehub`. * **Groups fatalities** by year and region to show where and when things escalated. * **Compares the top 2 violent actors** in a specific country (e.g., Military vs. Rebels). * **Breaks down event types** (Battles vs. Violence against civilians) to see tactical shifts. * **Advanced Mapping:** Uses `MarkerCluster` to handle high-density data and prevents "Mega-Events" by grouping on a daily level (`EVENT_DATE`) instead of just yearly. ## Technical Improvements * **Granular Aggregation:** Grouped data by coordinates and exact dates to ensure map markers represent specific incidents rather than inflated yearly totals. * **Date Parsing:** Implemented `dayfirst=True` and `errors='coerce'` to handle diverse date formats without crashing the script. * **Marker Clustering:** Integrated `folium.plugins.MarkerCluster` to keep the UI clean when analyzing hundreds of fatal events. ## Project Structure * `analysis.py`: Contains the logic, data cleaning, and processing functions. * `main.py`: The entry point. Interactive CLI for country selection and report generation. * `maps/`: Directory where the interactive HTML maps are stored. ## How to run Make sure you have pandas, folium, and kagglehub installed: ```bash pip install pandas folium kagglehub