EDA | Ford Car Price Prediction

What is EDA?

Exploratory Data Analysis (EDA) is the critical first step in any data science project. Before building models, we must understand the data — its structure, distributions, relationships, and quirks.

For this project, we conducted comprehensive EDA with 22 visualizations across multiple analysis types to extract actionable insights about used Ford vehicle pricing.

Analysis Types

Categorical Distributions

Count plots for transmission, fuel type, and model reveal market composition:

Manual transmission dominates the dataset
Petrol is the most common fuel type by far
Fiesta and Focus are the most listed models

Price Distribution

Histogram analysis shows:

Right-skewed distribution (mean > median)
Mean: £12,280 | Median: £11,291
Premium vehicles (Mustang) pull the mean higher
Log transformation brings distribution closer to normal

Correlation Heatmap

Pearson correlation reveals:

year — strong positive (newer = pricier)
mileage — moderate negative (more miles = cheaper)
engineSize — moderate positive
mpg — weak negative (efficient = budget models)

Outlier Detection

IQR method identifies:

Price: High-end outliers (Mustang, premium trims)
Mileage: Upper-end outliers (high-mileage commercial use)
MPG: Extreme outlier at 201.8 (physically implausible)

Pairplot

Pairwise relationships show:

Non-linear patterns between mileage and price
Clear clusters by vehicle type
Strong diagonal patterns in key features
2,000-record sample for performance

Violin Plots

Distribution by category:

Hybrid/Electric command premium prices
Automatic/Semi-Auto fetch higher prices than Manual
Wider price spread in premium categories
Clear median differences across groups

Average Price by Model

Model-level pricing hierarchy:

Mustang, Tourneo Custom, Ranger — highest average
Ka, Fiesta — budget end
Model type is a strong price driver
Clear stratification across Ford lineup

Boxplots & Scatter

Additional visualizations:

Year vs Price: Prices rise steadily with newer years
Engine Size vs Price: Larger engines = higher prices
Mileage vs Price: Clear negative trend
MPG vs Price: Weak negative relationship

Key EDA Insights

Top Positive Price Drivers

Year of registration — strongest predictor (newer = expensive)
Engine size — larger displacement = premium pricing
Model type — Mustang far outprices Ka and Fiesta
Fuel type — Hybrid and Electric command premiums
Transmission — Automatic costs more than Manual

Top Negative Price Drivers

Mileage — each additional mile reduces value
MPG — high efficiency often indicates budget models
Older vehicles — depreciation over time

Data Quality Observations

Year = 2060 outliers indicate data entry errors
MPG > 150 is physically implausible
Engine size = 0 likely represents missing data
154 duplicate records identified
Zero missing values — excellent data completeness

Statistical Findings

Price distribution is right-skewed (skewness ≈ +2)
Most common vehicle: Fiesta (budget-friendly)
Premium segment (Mustang, Ranger) is underrepresented
Manual transmission dominates the dataset
Petrol is by far the most common fuel type

Exploratory Data Analysis