How hudaa0412/mlops_telecom_customer_churn_prediction_model works
This is an incomplete, stage-oriented churn-ML template. The shown runtime starts in main.py, reads the raw CSV, and preprocesses it; the README documents later training, evaluation, prediction, registry, and serving capabilities. The model-stage files are present but empty in the excerpts, and the documented API module is absent from the tree, so those links reflect README claims rather than inspected implementation. The graph omits the API node because no matching source path is present.
Main entry points: main.py Pipeline entry · data_ingestion.py Data ingestion · data_preprocessing.py Preprocessing · processed Processed data
Components 16
ML workflow
- Pipeline entry
main.py - Data ingestion
src/data_ingestion.py - Preprocessing
src/data_preprocessing.py - Model training
src/model_building.py - Model evaluation
src/model_evaluation.py - Prediction
src/model_prediction.py
Data and artifacts
- Pipeline config
config.yaml - Raw churn data
data/raw/churn.csv - Interim data
data/interim - Processed data
data/processed - Model registry
models - Evaluation reports
reports
User interfaces
- Web console
frontend/index.html - Container services
docker/docker-compose.yml
Other components
- Pipeline operator
- Browser user
Connections 16
- Pipeline operator → Pipeline entry: runs
README.md - Pipeline entry → Data ingestion: calls
main.py - Data ingestion → Raw churn data: reads
- Data ingestion → Interim data: writes
- Pipeline entry → Preprocessing: calls
main.py - Preprocessing → Interim data: reads
- Preprocessing → Processed data: writes
- Processed data → Model training: reads
README.md - Model training → Model registry: saves model
README.md - Processed data → Model evaluation: reads
README.md - Model registry → Model evaluation: loads model
README.md - Model evaluation → Evaluation reports: writes metrics
README.md - Model registry → Prediction: loads model
README.md - Browser user → Web console: uses
README.md - Pipeline operator → Container services: starts services
README.md - Pipeline entry → Pipeline config: uses settings