BastiDiesel/financial-inclusion-africa-ml

# Financial Inclusion in Africa – Machine Learning Project ## Project Overview This project was developed as part of the Data Science & AI Engineering Bootcamp. The objective is to predict whether a respondent owns or uses a bank account based on demographic and socio-economic information. The project is based on the **Zindi Financial Inclusion in Africa** competition. --- ## Project Goal Financial inclusion plays an important role in economic development. The goal of this project is to develop a supervised machine learning model that predicts bank account ownership using survey data collected in four African countries. --- ## Dataset The dataset contains information from respondents in - Kenya - Rwanda - Tanzania - Uganda Each observation includes features such as - Country - Age - Household Size - Cellphone Access - Education Level - Job Type - Marital Status - Relationship to Household Head - Location Type The target variable is ```text bank_account ``` with the classes ```text Yes No ``` --- ## Project Workflow The notebook follows a complete end-to-end machine learning workflow. ### 1. Data Loading - Import datasets - Initial inspection - Missing value analysis ### 2. Exploratory Data Analysis (EDA) Analysis of - Target distribution - Cellphone access - Education level - Location type - Age - Country - Household size ### 3. Data Preparation - Target encoding - One-Hot Encoding - Train/Test Split - Preprocessing Pipeline ### 4. Machine Learning The following models were evaluated: - Dummy Classifier (Baseline) - Logistic Regression - Decision Tree - Random Forest ### 5. Model Evaluation Evaluation metrics include - Accuracy - Mean Absolute Error (MAE) - Precision - Recall - F1 Score - Confusion Matrix - Cross Validation ### 6. Model Interpretation The Random Forest model was further analysed using - Feature Importance - Grouped Feature Importance - Correlation Matrix ### 7. Final Model The best-performing Random Forest mode …