How kareem0088/tanzania-tourism-classification works
This is a tourism-cost classification project: its visible runtime code validates and cleans raw train/test data, engineers features, and provides categorical/target encoding intended for modeling. Configuration centralizes data paths, feature definitions, model parameters, and local MLflow tracking setup. The tree also shows feature datasets and a baseline model artifact, but no sampled code establishes how they are produced or consumed. Training, evaluation, and prediction modules are empty in the excerpts; notebook workflows are therefore not represented as verified runtime stages.
Main entry points: builder.py Feature engineering · preprocessing.py Data preprocessing · raw Raw tourism data · config.py Project configuration
Components 9
Data preparation
- Data preprocessing
src/preprocessing.py - Feature engineering
src/features/builder.py
Modeling support
- Categorical encoding
src/features/encoding.py - Project configuration
src/config.py - Baseline model artifact
models/baseline-model.json
Experiment tracking
- MLflow SQLite store
mlflow.db
Other components
- Project user
- Raw tourism data
data/raw - Feature datasets
data/features
Connections 5
- Project user → Data preprocessing: runs
- Data preprocessing → Raw tourism data: loads
src/preprocessing.py - Feature engineering → Raw tourism data: loads
src/features/builder.py - Feature engineering → Feature datasets: writes
src/features/builder.py - Project configuration → MLflow SQLite store: connects
src/config.py