samsonafo / cars_software Goto Github PK

This Project is aimed at building a predictive model for predicting used-car prices in Nigeria.

The Poperties Being considered are: 1. Make 2. Year 3. Model 4. Mileage 5. Transmission 6. Color 7. Location 8. History 9. Car-Rating

This file scraps data from the cars45 website. copy the webpage of the car, you want to scrap into cell 3. Keep copying the link and run only cell 3 and 4. Be careful not to run cell 2, this will re-initialize the dataframe.

This file scraps data from the cheki website. copy the webpage of the car, you want to scrap into cell 3. Keep copying the link and run only cell 3 and 4. Be careful not to run cell 2, this will re-initialize the dataframe. Happy Scraping.

This file scraps data from car from autochek which is also redirecting the datas from Cheki.com.ng. In this file We scraped 14,195 rows of cars and saved into a CSV file afterwards.

For model development - Here I compared the performance of different models. (Linear Regression, k-Nearest Neigbours, Random Forest and Gradient Boosted Trees) with missing rating set to 2.0

For model development - Here I used GridSearchCV to find the best parameters for the Gradient Boosted Trees and K-Nearest Neigbhours. (The 2 best performing models from model_2.ipynb)

For model development - Here I compared the performance of different models. (Linear Regression, k-Nearest Neigbours, Random Forest and Gradient Boosted Trees) with missing rating set to 3.0