🌌 The CosmoClassifier

A machine learning–powered web application that classifies celestial objects (Galaxies, Stars, and Quasars) using data from the Sloan Digital Sky Survey DR18 (SDSS DR18).
Built with Flask, Scikit-Learn, and Pandas, this project blends astronomy with machine learning to create an interactive and educational tool for space data analysis.
This project demonstrates the intersection of astrophysics and applied machine learning which has made space data exploration more accessible.

---

🚀 Overview

The CosmoClassifier uses a trained Random Forest Classifier optimized through Sequential Feature Selection (SFS) to predict whether a given object in the universe is a Galaxy, Star, or Quasar (QSO).
The model is hosted via a Flask web app where users can input photometric parameters (like magnitudes and redshift) and get instant predictions with probability scores.

---

🧠 Motivation

The Sloan Digital Sky Survey (SDSS) is one of the largest astronomical databases ever created. But handling its raw data isn’t exactly beginner-friendly.
This project simplifies that complexity—transforming high-dimensional SDSS data into an approachable, real-time classification tool.

---

📸 Screenshots

---

📊 Dataset

• Source: Sloan Digital Sky Survey DR18 – Kaggle Dataset
• Classes:
  • GALAXY → 0
  • STAR → 1
  • QSO → 2
• Feature Engineering:
  • Created new color-contrast features:
    • u_g_color = u − g
    • g_r_color = g − r
    • r_i_color = r − i
    • i_z_color = i − z

---

⚙️ Model Architecture

Step	Description
Imputation	Missing values handled using median strategy
Scaling	Standardized with StandardScaler
Feature Selection	Sequential Feature Selector (SFS) with Random Forest
Classifier	Random Forest Classifier (max_depth = 10, n_estimators = 150, class_weight = "balanced")

Final model artifacts are serialized with joblib as:

models/
├── pipe.pkl
└── column_names.pkl

---

🧪 Model Selection Research

Previously, multiple combinations of models and dimensionality-reduction methods were tested before finalizing the architecture. This was a manual process which is now replaced by scikit-learn's RandomizedSearchCV algorithm, which allows us to find the best classifier and hyperparameter setting.

Here is the Classification Report which was calculated after fitting the best model discovered by RandomizedSearchCV:

Class	Precision	Recall	F1-Score	Support
0 (GALAXY)	0.99	0.99	0.99	10469
1 (STAR)	0.99	1.00	0.99	7446
2 (QSO)	0.97	0.96	0.97	2085
Accuracy	-	-	0.99	20000
Macro Avg	0.98	0.98	0.98	20000
Weighted Avg	0.99	0.99	0.99	20000

You can find the code for this in research.ipynb file included in the repo. Note that this result is for the full dataset with all original features.

---

🧩 Project Structure

THE_COSMOCLASSIFIER/
├── Datasets/
│   └── SDSS_DR18.csv        # Primary dataset for training and analysis
├── models/
│   ├── column_names.pkl     # List of column names used in the model
│   └── pipe.pkl             # Serialized machine learning pipeline (model)
├── notebooks/
│   ├── research_2.ipynb     # Research notebook used later for implementing feature reduction 
│   ├── research_LR.ipynb    # Research notebook for Logistic Regression
│   ├── research_RF.ipynb    # Research notebook for Random Forest
│   └── research_SVC.ipynb   # Research notebook for Support Vector Classifier
├── static/                  # Static assets for the web application
│   ├── script.js            # Frontend JavaScript logic
│   └── style.css            # Frontend CSS styling
├── templates/
│   └── index.html           # HTML template for the web application
|
├── .gitignore               # Files and directories to be excluded from version control
├── app.py                   # Main Flask application file for the web service
├── fit.py                   # Script for training and saving the machine learning model
├── LICENSE                  # Licensing information
└── requirements.txt         # Python package dependencies

---

💻 Installation & Usage

1️⃣ Clone the Repository

git clone https://github.com/ByteBard58/The_CosmoClassifier
cd The_CosmoClassifier

2️⃣ Install Dependencies

pip install -r requirements.txt

3️⃣ Run the App

python app.py

Then open your browser at http://127.0.0.1:5000/

If the serialized model isn’t found, app.py automatically trains and saves a new one using fit.py.

---

🐳 Run the app directly via Dockerhub Image

A Dockerhub repository is created where I have pushed the docker image which contains the entire Cosmoclassifier app.

The image is built on both ARM64 and AMD64 architectures, so that it can run on almost all major computers. You can run the app easily by using the Dockerhub Image. Here's how you can do it:

1. Install Docker Desktop and sign-in. Make sure the app is functioning properly.

2. Open Terminal and run:

docker pull bytebard101/cosmoclassifier:latest
docker run --rm -p 5000:5000 bytebard101/cosmoclassifier:latest

3. If your machine faces a port conflict, you will need to assign another port. Try to run this:

docker run --rm -p 5001:5000 bytebard101/cosmoclassifier:latest

If you followed Step 2 and the command ran successfully, then DO NOT follow this step.

4. The app will be live at localhost:5000. Open your browser and navigate to http://127.0.0.1:5000 (or http://127.0.0.1:5001 if you followed Step 3).

Check Docker Documentation to learn more about Docker and it's commands.

---

🌠 Web Interface

Users can input photometric parameters such as:

• Right Ascension (deg)

• Declination (deg)

• Redshift

• PSF Magnitude (r band)

• Color contrast features (u_g_color, g_r_color, r_i_color, i_z_color)

The app returns the predicted class with its confidence probabilities.

Note: Due to the absence of front-end web development expertise and the lack of other contributors for the project, I had to resort to using AI tools (such as LLM services like ChatGPT™, Grok™, and GitHub Copilot™) to create a sophisticated front-end for the web app.

---

🧰 Tech Stack

• Languages: Python, HTML, CSS, JavaScript

• Libraries: Flask, Scikit-Learn, Pandas, NumPy, Joblib, Notebook

• Dataset Source: Sloan Digital Sky Survey DR18 (Kaggle)

---

🪐 Author

Sakib ( ByteBard58 )

Student | Aspiring Computer Engineer | AI & ML Enthusiast

📍 GitHub Profile: ByteBard58

---

😃 Appreciation

Thank You for taking the time to review my work. I hope you enjoyed it and found it interesting. It would mean a lot to me if you could star it on GitHub 🌟

If you have any questions, suggestions, or anything you’d like to discuss, please don’t hesitate to reach out. You can find my contact information on my GitHub profile page. I’m all ears! 😊

Have a great day !