Ford Car Price Prediction Model

Overview

A machine learning project to predict Ford car prices using regression analysis with an accuracy of 82%.

Dataset

• Source: ford.csv
• Features: year, mileage, tax, mpg, engineSize, model, transmission, fuelType
• Target: price

Project Workflow

1. Exploratory Data Analysis (EDA)

• Analyzed dataset shape and structure
• Examined data types and statistical summaries
• Verified data quality (no null values)
• Generated correlation heatmap

2. Data Cleaning

• Removed duplicate records
• Retained all original features

3. Data Visualization

• Histograms with KDE for numerical features
• Line plots for price trends by year and model
• Violin and box plots for categorical analysis
• Correlation heatmap

4. Feature Engineering

• One-Hot Encoding: Converted categorical variables (model, transmission, fuelType)
• Feature Scaling: Applied StandardScaler to numerical features (mileage, tax, mpg, engineSize, year)

5. Data Splitting

• Training set: 80%
• Testing set: 20%
• Random state: 42

6. Model Training

• Algorithm: Linear Regression
• Successfully trained on preprocessed features

7. Model Evaluation

• R² Score: 0.82 (82% accuracy)
• Adjusted R² Score: Calculated based on test set dimensions

Results

The model explains 82% of the variance in Ford car prices, demonstrating strong predictive performance.