• A web app that estimates the selling price of the used cars to help customers to negotiate from dealers.
• Optimized Random Forest Regressor using RandomizedsearchCV to reach the best model.
• Built a client facing API using flask and deployed into heroku cloud.
• Demo: https://sell-my-car.herokuapp.com/

Python Version: 3.7

Packages: pandas, numpy, sklearn, matplotlib, seaborn, flask, pickle

For Web Framework Requirements: pip install -r requirements.txt

• First, I transformed the categorical variables into dummy variables. I also split the data into train and tests sets with a test size of 20%.
• I used ExtraTreeRegressor to know the feature importance.

• I tried Random Forest model and evaluated them using Negative mean squared error and tuned the model using randomized SearchCV to obtain best results.

The model performed very well.

In this step, I built a flask API endpoint that was hosted on a local webserver and deployed into heroku cloud. The API endpoint takes in a request with a list of values from the user and returns estimated selling price.