How anjali22-lgtm/telecom-churn-prediction works
This is a Streamlit dashboard for predicting telecom customer churn from manually entered customer attributes. Its central workflow collects inputs, preprocesses them, scores them with a saved model, classifies risk, and presents prediction results, customer summary, and visualizations. The README also describes the broader data-to-model training workflow, but the repository tree exposes no source modules for those stages. Relationships below are limited to the sampled app excerpt and documented capabilities; no training pipeline or persistent data store is shown.
Main entry points: app.py Streamlit app · features.pkl Feature schema · scaler.pkl Feature scaler · churn_model.pkl Churn model
Components 15
Dashboard
- Streamlit app
app.py - Customer inputs
app.py - Lottie animation
app.py
Prediction
- Input preprocessing
app.py - Feature schema
features.pkl - Feature scaler
scaler.pkl - Churn model
churn_model.pkl
Results
- Risk classification
app.py - Customer summary
app.py - Probability gauge
app.py - Probability pie chart
app.py - Model metrics
app.py - Prediction dashboard
app.py
Other components
- Customer
- Animation service
Connections 14
- Customer → Customer inputs: enters attributes
app.py - Streamlit app → Lottie animation: renders
app.py - Lottie animation → Animation service: fetches animation
app.py - Customer inputs → Input preprocessing: prepares input
app.py - Input preprocessing → Feature schema: orders columns
app.py - Input preprocessing → Feature scaler: scales values
app.py - Streamlit app → Churn model: loads model
app.py - Streamlit app → Input preprocessing: preprocesses input
app.py - Streamlit app → Risk classification: classifies probability
app.py - Streamlit app → Prediction dashboard: renders results
app.py - Prediction dashboard → Customer summary: shows summary
app.py - Prediction dashboard → Probability gauge: shows probability
app.py - Prediction dashboard → Probability pie chart: shows distribution
app.py - Prediction dashboard → Model metrics: shows metrics
app.py