How pranathir73/indian-pre-owned-car-price-prediction works
This is a Streamlit app for estimating Indian pre-owned car prices. A user supplies vehicle details for an instant valuation, or uploads a CSV for batch predictions; both paths engineer features and call the saved model, then show prices with ±5% bands. The app also presents pricing trends from the training CSV. The sampled app.py shows these flows and the model/data reads; the notebook and training process were not inspected, so model creation is not mapped.
Main entry points: app.py Batch CSV upload · best_car_price_model.joblib Saved price model · feature_metadata.json Feature metadata · Cap_Training_Data_2025.csv Training data
Components 14
User interface
- Vehicle form
app.py - Batch CSV upload
app.py
Prediction
- Model and metadata loader
app.py - Saved price model
best_car_price_model.joblib - Feature metadata
feature_metadata.json - Feature engineering
app.py - Single-car valuation
app.py - Batch prediction
app.py - Price bands and warnings
app.py - Prediction CSV download
app.py
Market insights
- Training data loader
app.py - Training data
Cap_Training_Data_2025.csv - Pricing trend charts
app.py
Other components
- User
Connections 15
- User → Vehicle form: enters details
app.py - User → Batch CSV upload: uploads CSV
app.py - Model and metadata loader → Saved price model: loads model
app.py - Model and metadata loader → Feature metadata: loads metadata
app.py - Vehicle form → Feature engineering: engineers features
app.py - Feature engineering → Single-car valuation: returns features
app.py - Single-car valuation → Saved price model: predicts price
app.py - Single-car valuation → Price bands and warnings: computes bands
app.py - Batch CSV upload → Feature engineering: engineers features
app.py - Batch CSV upload → Saved price model: predicts prices
app.py - Batch CSV upload → Prediction CSV download: builds CSV
app.py - Single-car valuation → Feature metadata: reads options and metrics
app.py - Training data loader → Training data: reads data
app.py - Training data loader → Pricing trend charts: supplies data
app.py - User → Prediction CSV download: downloads results
app.py