ann-mw/Kenya-Crime-Prediction-ARIMAX-

# Kenya-Crime-Prediction-ARIMAX- This project uses the ARIMAX model to forecast crime rates across Kenyan counties from 2019 to 2023. It leverages external socio-economic and environmental factors to enhance prediction accuracy. > 🔍 Built for academic research and statistical modeling purposes. --- ## 📈 Project Overview The goal of this project is to predict future crime trends in Kenya using ARIMA and ARIMAX time series models. The study incorporates external variables such as: - Poverty rates - Unemployment levels - Urban population growth - Education access --- ## 🔧 Technologies Used - **Python** - `pandas`, `numpy`, `matplotlib`, `seaborn` - `statsmodels` (for ARIMA & ARIMAX) - `sklearn` (metrics evaluation) - Jupyter Notebook --- ## 🔍 Methodology 1. **Exploratory Data Analysis (EDA)** Cleaning, visualization, and trend detection. 2. **Modeling** - ARIMA: Baseline crime prediction - ARIMAX: Enhanced forecasting using external variables 3. **Evaluation** - MAE (Mean Absolute Error) - RMSE (Root Mean Squared Error) --- ## 📊 Results & Insights - ARIMAX outperformed ARIMA in predictive accuracy. - Strong correlation found between poverty & crime in urban counties. - Seasonal patterns observed in theft and assault crimes.