🚗 Vehicle Market Segmentation – Clustering Analysis

Quick Snapshot

• Dataset: Vehicle specifications (156 records, 8 numerical features)
• Task: Market segmentation (unsupervised learning / clustering)
• Techniques compared: Hierarchical Clustering, K-Means
• Best clustering: Hierarchical Clustering (Ward / Average Linkage)
• Key insight: Vehicles naturally group by engine size, horsepower, weight, and dimensions

📌 Project Overview

This project applies unsupervised machine learning clustering techniques to segment vehicles based on technical specifications.

Note: The dataset was provided as a Data Science Bootcamp task and does not represent any real vehicle market data. This is purely an educational exercise.

Objective:

• Discover natural vehicle clusters
• Compare Hierarchical Clustering vs K-Means
• Determine optimal number of clusters
• Translate clustering results into business insights

---

🎯 Background

1. Segment vehicles using technical specs: engine, horsepower, weight, size, fuel efficiency
2. Compare multiple clustering algorithms using silhouette score and cluster interpretability
3. Identify which approach produces the most meaningful and balanced clusters
4. Support hypothetical business decisions, like product positioning or prototype comparison

---

🗂️ Dataset Overview

Source: Provided by Data Science Bootcamp (156 records)

Features Used for Clustering:

Feature	Description
engine_s	Engine size (liters)
horsepow	Horsepower
wheelbas	Wheelbase (inches)
width	Vehicle width (inches)
length	Vehicle length (inches)
curb_wgt	Curb weight (1,000 lbs)
fuel_cap	Fuel capacity (gallons)
mpg	Miles per gallon

Problem Type: Unsupervised learning - Clustering

⚠️ Disclaimer Dataset is synthetic/educational; any real-world conclusions should not be drawn from this data.

---

🛠️ Tools I Used

Programming & Libraries:

• Python – Analysis and model implementation
• NumPy – Numerical computations and array operations
• Pandas – Data manipulation and analysis
• Matplotlib / Seaborn / SciPy – Data visualization and dendrograms
• Scikit-learn – Machine learning models, preprocessing, metrics and model evaluation

Development Environment:

• Jupyter Notebook – Interactive development and analysis
• Google Colab – Optional cloud-based notebook environment

---

🚀 Getting Started

Prerequisites

• Python 3.11 or higher
• Jupyter Notebook or Google Colab
• Conda (recommended) or pip

Installation

1. Clone the repository

    git clone https://github.com/NadiaRozman/ML_Clustering_Vehicle_Segmentation.git
    cd ML_Clustering_Vehicle_Segmentation

2. Install dependencies

   Option 1: Using Conda

   conda env create -f environment.yml
   conda activate ml_vehicle_clustering

   Option 2: Using pip

   pip install -r requirements.txt

3. Launch Jupyter Notebook

   jupyter notebook

4. Open and run the notebook

   • Navigate to notebook/Vehicle_Market_Segmentation_Clustering.ipynb
   • Ensure vehicle_specifications.csv is in the correct directory
   • Run all cells to reproduce the analysis

---

🔬 Clustering Models Implemented

1. Hierarchical Clustering

• Linkage methods compared: Ward, Complete, Average
• Evaluated optimal number of clusters using dendrograms and silhouette scores

Ward Linkage Dendrogram

Figure 1: Hierarchical clustering using Ward linkage minimizes within-cluster variance.

Complete Linkage Dendrogram

Figure 2: Hierarchical clustering using Complete linkage minimizes the maximum distance between clusters.

Average Linkage Dendrogram

Figure 3: Hierarchical clustering using Average linkage minimizes the average distance between all pairs.

Hierarchical Clustering Silhouette Scores

Figure 4: Silhouette scores for hierarchical clustering methods; helps identify optimal number of clusters.

2. K-Means Clustering

• Tested clusters from k=2 to k=10
• Optimal k selected based on silhouette score and inertia (WCSS)

Elbow Method & Silhouette Scores

Figure 5: Elbow method and silhouette scores for K-Means; optimal k = 2.

Final K-Means Cluster Visualization

Figure 6: K-Means clustering scatter plot showing 2 vehicle clusters based on features.

---

Cluster Analysis

Sample Vehicles by Cluster (K-Means, k=2)

Cluster	Sample Models	Avg Engine	Avg HP	Avg MPG	Avg Weight
0	Acura Integra, Audi A4, BMW 323i	2.3 L	150	26.5	2.88k lbs
1	Acura TL, Audi A6, Buick Regal	3.7 L	218	21.5	3.82k lbs

Clusters differentiate vehicles roughly by engine size, horsepower, weight, and dimensions.

---

Hierarchical vs K-Means

Metric	Hierarchical (Average)	K-Means
Clusters	2	2
Silhouette Score	0.5147	0.3666

🏆 Winner: Hierarchical Clustering (Average Linkage) – higher silhouette score, though K-Means produced more balanced clusters.

---

🔍 Insights & Business Implications

• Vehicles naturally segment into low-power / compact vs high-power / large groups
• Hierarchical Clustering captures outliers more clearly
• K-Means provides balanced clusters suitable for portfolio comparison
• Hypothetical application: guiding prototype design or market positioning

---

📚 What I Learned

• Applied multiple clustering algorithms and compared results
• Learned feature scaling importance for distance-based methods
• Visualized hierarchical relationships with dendrograms
• Interpreted clusters with silhouette score and sample data
• Translated unsupervised learning outputs into actionable insights

---

🔮 Future Enhancements

• Explore dimensionality reduction (PCA, t-SNE) for better visualization
• Hyperparameter tuning for K-Means initialization
• Include more features or categorical variables
• Deployment: dashboard visualization (Streamlit / Plotly)
• Cross-validation for unsupervised clustering stability

---

✨ Created by Nadia Rozman | January 2026

📂 Project Structure

ML_Clustering_Vehicle_Segmentation/
│
├── data/
│   └── vehicle_specifications.csv
│
├── images/
│   ├── 1_dendrogram_ward.png
│   ├── 2_dendrogram_complete.png
│   ├── 3_dendrogram_average.png
|   ├── 4_hierarchical_silhouette.png
│   ├── 5_kmeans_elbow_silhouette.png
│   └── 6_kmeans_scatter.png
│
├── notebooks/
│   └── Vehicle_Market_Segmentation_Clustering.ipynb
│
├── requirements.txt
├── environment.yml
└── README.md

🔗 Connect with me

• GitHub: @NadiaRozman
• LinkedIn: Nadia Rozman

⭐ If you found this project helpful, please consider giving it a star!