mah-trigui/financial-inclusion-in-africa-Zindi-Competition

# Financial Inclusion in Africa — Bank Account Prediction Pipeline This competition is hosted on Zindi, a machine learning platform for data science challenges. Here is the link to the competition: Financial Inclusion in Africa 🌾 - Knowledge Ranked in the TOP 7% --- Organized classification pipeline · Binary prediction · Multiple ML models with ensemble. --- ## Competition Overview | Item | Details | |---|---| | Task | Predict whether an individual has or uses a **bank account** | | Target | Binary classification: Yes / No | | Geography | Multiple African countries (Kenya, Rwanda, Tanzania, Uganda) | | Features | Demographics, socioeconomic, household, access indicators | | Evaluation | Likely accuracy or ROC-AUC (standard binary classification) | | Submission format | One row per `uniqueid` with predicted probabilities | --- ## Dataset | File | Rows | Columns | Description | |---|---|---|---| | `Train_v2.csv` | ~33,000 | 10+ | Training set with target `bank_account` | | `Test_v2.csv` | ~14,000 | 9+ | Test set — `bank_account` withheld | **Key Features:** - **Geography:** `country`, `location_type` (Rural/Urban) - **Demographics:** `age_of_respondent`, `gender_of_respondent`, `marital_status`, `relationship_with_head` - **Socioeconomic:** `job_type`, `education_level`, `household_size` - **Access:** `cellphone_access` - **Target:** `bank_account` (Yes/No) --- ## Pipeline Structure ``` pipeline/ ├── 00_config.r # Libraries, paths, constants, seeds, utilities ├── 01_data_loading.r # Load Train_v2.csv and Test_v2.csv ├── 02_data_cleaning.r # Remove inconsistencies, impute, consolidate ├── 03_feature_engineering.r # Binning, interactions (age, income, geo, demo) ├── 04_feature_selection.r # Boruta, RFE, correlation analysis ├── 05_preprocessing.r # Encoding, scaling, prepare for models ├── 06_models.r # Train RF, Ranger, XGBoost, GLM, SVM, NB, GBM ├── 07_evaluation.r # Model leaderboard …