https://github.com/zsxkib/most-under-and-over-priced-cars

Determine what influences and drives car prices given technical specs and identify which car(s) are the most under/overpriced and why.
https://github.com/zsxkib/most-under-and-over-priced-cars

analysis cars-dataset explainable-artificial-intelligence interpretable-machine-learning outlier-detection shap xgboost-regression

Last synced: 16 days ago
JSON representation

Determine what influences and drives car prices given technical specs and identify which car(s) are the most under/overpriced and why.

• Host: GitHub
• URL: https://github.com/zsxkib/most-under-and-over-priced-cars
• Owner: zsxkib
• Created: 2021-09-22T14:34:39.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2021-09-22T15:16:03.000Z (over 3 years ago)
• Last Synced: 2024-12-16T11:06:00.425Z (18 days ago)
• Topics: analysis, cars-dataset, explainable-artificial-intelligence, interpretable-machine-learning, outlier-detection, shap, xgboost-regression
• Language: Jupyter Notebook
• Homepage:
• Size: 3.84 MB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Most Under/Over Priced Cars

#### The task is to determine what influences and drives the car prices given its technical specs. Specifically, which car(s) I believe to be most under/overpriced and why.

## Overview

_**This was a take-home assignment for a Graduate Data Analyst Consultant position at a Machine Learning Consultancy**_

Within the repo you will find the following:

*	car_data_cleaned.csv 

> (Raw data .csv file sent by the Machine Learning Consultancy)

*	code.(pdf/html)

> PDF (.pdf) & Web (.html) versions of my notebook, just in case there’s technical difficulties with original Jupyter notebook)

*	code.ipynb 

> (Jupyter Notebook containing all analysis, code, plots, etc)

*	method_results_conclusion.pptx 

> Three slide presentation outlining method, results, and conclusion)

## Method

* Exploratory Data Analysis

* Clean Data

* XGBoost Regressor with 5-Fold Cross-Validation

* Under/Over-priced Car Ranking per Fold

* XGBoost Regressor on Top/Bottom 50 in Ranking

* SHapley Additive exPlanations (SHAP)

## Results

### Top 3 Most Over-Priced Cars:

![image](/plots/most_over_priced_car_and_why.png?raw=true "Top 3 Most Over-Priced Cars SHAP Force Plot")

* **2018 Mercedes-Benz G Class Specs: G550 4x4 Squared**

> True MSRP = 227 300

> Predicted MSRP = 120 390.77

> Ratio (True/Predicted) = 1.888

* 2019 Jaguar XE SV Specs: Project 8 AWD

> True MSRP = 187 500

> Predicted MSRP = 119 680.59

> Ratio (True/Predicted) = 1.566

* 2018 Porsche 911 Specs: GT2 RS Coupe

>True MSRP = 293 200

>Predicted MSRP = 193 417.28

> Ratio (True/Predicted) = 1.515

### Top 3 Most Under-Priced Cars

![image](/plots/most_under_priced_car_and_why.png?raw=true "Top 3 Most Under-Priced Cars SHAP Force Plot")

* **2018 Jeep Grand Cherokee Specs: Altitude 4x4**

> True MSRP = 33 195

> Predicted MSRP = 47 957.96

> Ratio (True/Predicted) = 0.692

* 2018 Jeep Grand Cherokee Specs: Upland 4x4

> True MSRP = 33 195

> Predicted MSRP = 47 957.96

> Ratio (True/Predicted) = 0.692

* 2019 Dodge Challenger Specs: R/T Scat Pack Widebody RWD

> True MSRP = 38 995

> Predicted MSRP = 56 335.17

> Ratio (True/Predicted) = 0.692

![image](/plots/why.png?raw=true "Most Under/Over-Priced Cars SHAP Waterfall Plots")