Price-Foreteller-For-Preowned-Cars

A machine learning based prediction model that is used to predict the price of used cars based on data given by user.

Modules used

For ML model creation

• Anaconda - (Jupyter Notebook,Scikit,Pandas,Python3,Numpy)
• LightGBM

For creating website

• Flask
• requests
• jsonify
• Pickle
• HTML5-CSS

Label encoding on features

Fuel Type

• 0 - CNG
• 1 - Diesel
• 2 - Petrol
• 3 - LPG
• 4 - Electric

Brand Name

• 0 - Ambassador
• 1 - Audi
• 2 - BMW
• 3 - Bentley
• 4 - Chevrolet
• 5 - Datsun
• 6 - Fiat
• 7 - Force
• 8 - Ford
• 9 - Hindustan
• 10 - Honda
• 11 - Hyundai
• 12 - ISUZU
• 13 - Jaguar
• 14 - Jeep
• 15 - Lamborgini
• 16 - LandRover
• 17 - Mahindra
• 18 - Maruti
• 19 - Mercedes-Benz
• 20 - MiniCooper
• 21 - Mitsubishi
• 22 - Nissan
• 23 - OpelCorsa
• 24 - Porshe
• 25 - Renault
• 26 - Skoda
• 27 - Smart
• 28 - Tata
• 29 - Toyota
• 30 - Volkswagen
• 31 - Volvo

Location

• 0 - Mumbai
• 1 - Pune
• 2 - Chennai
• 3 - Coimbatore
• 4 - Hyderabad
• 5 - Jaipur
• 6 - Kochi
• 7 - Kolkata
• 8 - Delhi
• 9 - Bangalore
• 10 -Ahmedabad

Transmission

• 0 - Manual
• 1 - Automatic

Owner Type

• 0 - First
• 1 - Second
• 2 - Fourth & Above
• 3 - Third

Steps to run .ipynb files

• Download & Install Anaconda
• Use Jupyter Notebook to run the file
• OR
• Use Google Colab to run .ipynb files in the cloud.

Steps for local deployment

• Open cmd
• Change directory to the place where the code is saved
• run "python app.py" to deploy locally.

Note:-

• If changes in dataset is made, model might give out undesirable outputs.
• Install all python modules mentioned above before running the project.
• Output.xlsx gives output after prediction based on test data.
• While using the website, try to avoid giving vague inputs as it provides wrong results.

The site is live.

Video Narration for the project is available here.