Data-Visualization-project

This project explores the Breast Cancer Wisconsin (Diagnostic) dataset from the UCI Machine Learning Repository. The aim is to apply a range of visualisation techniques to better understand the dataset, highlight patterns, and distinguish between benign and malignant diagnoses.

---

Project Overview

The project demonstrates how visualisation can support exploratory data analysis by:

• Highlighting correlations between features.
• Showing class separability (benign vs malignant) across multiple attributes.
• Identifying trends, outliers, and redundancies in the dataset.

The workflow includes:

1. Heatmap (Correlation Plot) – Uses Pearson correlation to examine relationships between features.
2. Beeswarm Plot (Quasirandom) – Helps avoid overplotting and makes the separation between benign and malignant cases clearer.
3. Histogram – Displays feature distributions, highlighting which features best separate diagnoses.
4. Scatter Plots with Ellipses – Provides confidence regions to show overlap or separability between classes.

---

Dataset

The dataset contains diagnostic features of breast cancer cases, with each feature having three variations:

• Mean
• Standard error (SE)
• Worst

The target variable indicates whether a case is Benign (B) or Malignant (M).

---

Visualisations

• Heatmap: Shows the strength and direction of correlations between features.
• Beeswarm plots: Display individual data points without overlap, aiding visual comparison across 30 features.
• Histograms: Reveal variance within features and show how well they separate diagnoses.
• Scatter plots with ellipses: Illustrate 95% confidence regions to highlight areas of overlap and separation between classes.

---

Key Findings

• Visualisation helps in identifying redundant features with high correlation.
• Certain features clearly separate benign and malignant cases, while others show heavy overlap.
• Scatter plots with confidence ellipses provide a more nuanced view of class separability.

---

Technologies Used

• Python
• Matplotlib / Seaborn
• Pandas / NumPy

---

How to Run

1. Clone this repository:

   git clone https://github.com/yourusername/breast-cancer-visualisation.git
   cd breast-cancer-visualisation