🌌 CosmoLink

Hierarchical Grouping of Galaxies Based on Spatial Coordinates and Properties

---

🚀 Project Overview

CosmoLink is a machine learning project focused on clustering galaxies based on their spatial coordinates and astrophysical properties.
Using Agglomerative Hierarchical Clustering, CosmoLink uncovers natural groupings and structural patterns in galactic data obtained from the Sloan Digital Sky Survey (SDSS).

The project aims to reveal hidden cosmic structures by analyzing:

• Right Ascension (RA) and Declination (Dec) (spatial positions)
• Astrophysical properties (like redshift, magnitudes)

🛰️ Dataset

• Source: Sloan Digital Sky Survey (SDSS)
• Contents: Spatial coordinates, redshift values, photometric properties of galaxies.

🧠 Methodology

1. Data Loading:
   Load galaxy data from SDSS.

2. Preprocessing:

   • Handle missing values
   • Normalize spatial coordinates and features

3. Feature Selection:

   • Spatial coordinates (RA, Dec)
   • Redshift and other astrophysical properties

4. Clustering:

   • Apply Agglomerative Hierarchical Clustering
   • Choose optimal linkage method (ward/complete/average)

5. Visualization:

   • Scatter plots colored by cluster labels
   • Dendrogram analysis

6. Evaluation:

   • Analyze clustering performance
   • Explore physical significance of identified groups

📊 Results

• Identified meaningful galactic clusters based on spatial distribution.
• Visualized cosmic structures and potential galaxy groupings.
• Explored hierarchical relationships between galaxy clusters.

Here's a 3D visualization of Galaxy Clusters:

✨ Future Work

• Incorporate more astrophysical features (spectral lines, mass estimates).
• Compare with other clustering techniques (e.g., DBSCAN, KMeans).
• 3D clustering using RA, Dec, and redshift as distance proxy.