How shivharerachit/autovaluator works
Project type and purpose
• AutoValuator is a lightweight web application whose goal is to predict used‐car prices.
• It consists of a static front end (HTML/JS) and a Python Flask backend that hosts a pre‐trained ML model.
• The “Model” folder also contains a Jupyter notebook and data for offline training, but at runtime only the pickle file and column metadata are required.
File structure
• Top level
• .gitignore – standard VCS exclusions
• README.md – project name (“AutoValuator”)
• Documentation/ – PDF reports (not part of runtime)
• FrontEnd/index.html – single‐page UI for collecting input features and displaying results
• Model/
– PredictCarPrice.pkl – serialized ML model
– column_names.txt – feature ordering info
– flask-api.py – Flask app exposing a “/predict” endpoint
– requirements.txt – Python deps (Flask, scikit‐learn, pandas, etc.)
– cardekho_dataset.csv & TrainingModel.ipynb – offline training pipeline
README
• The README only names the project but does not detail architecture. All architectural clues come from file names & structure.
Detail
• Essential runtime pieces—static front end, REST endpoint, model files.
• Avoid over-splitting; focus on the clear boundaries and interactions described above.