Multivariate Analysis: PCA, Factor Analysis, Clustering & MDS

Application of multivariate statistical techniques on real-world consumer datasets to uncover patterns, segmentation, and relationships.

---

Overview

This project applies advanced multivariate analysis methods to understand consumer behavior and product positioning.

Using real-world datasets, the analysis explores dimensionality reduction, latent factor identification, clustering, and similarity mapping to extract meaningful insights.

---

Techniques Used

Principal Component Analysis (PCA)

• Reduces dimensionality of high-dimensional data
• Identifies key components explaining maximum variance

Factor Analysis

• Extracts latent variables influencing observed data
• Helps interpret underlying consumer decision factors

Cluster Analysis (K-Means)

• Segments consumers into homogeneous groups
• Useful for market segmentation and targeting

Multidimensional Scaling (MDS)

• Visualizes similarity/dissimilarity between products
• Helps understand product positioning

---

Dataset

This project uses two real-world datasets:

• Survey Dataset (Survey.csv) Contains demographic and preference data of Bangalore home buyers, including income, budget, amenities, and decision factors. Used for PCA, Factor Analysis, and Cluster Analysis.

• Ice Cream Dataset (icecream.csv) Contains attributes and ratings of dairy brands such as price, taste, flavor, and shelf life. Used for Multidimensional Scaling (MDS).

---

Key Insights

• Consumer preferences are driven by a few dominant latent factors
• Clear segmentation exists among home buyers based on income and preferences
• PCA effectively reduces complexity while retaining key information
• MDS reveals positioning differences among competing brands

---

Applications

• Market segmentation and targeting
• Consumer behavior analysis
• Product positioning and branding
• Decision-making in real estate and FMCG sectors

---

Tech Stack

• Python (pandas, scikit-learn, matplotlib)
• R (stats, cluster, factoextra)
• Statistical modeling and machine learning techniques

---

Project Structure

• analysis.ipynb – Python implementation
• analysis.R – R implementation
• README.md – project overview
• project_report.pdf – detailed report
• data.csv – dataset