DJS NOVA - Technical AI Tasks

Status: All recruitment tasks completed ✅

Comprehensive solutions for DJS Nova Co-Comm recruitment tasks, featuring exploratory data analysis, machine learning, and astronomical data processing.

---

📋 Table of Contents

• About
• Repository Structure
• Completed Tasks
• Requirements
• Quick Start
• Usage
• Results & Evaluation
• Contributing
• License
• Contact

---

🎯 About

This repository contains complete solutions for DJS Nova technical AI tasks, demonstrating expertise in:

• Exploratory Data Analysis (EDA) on astronomical datasets (SDSS)
• Machine Learning for regression and classification
• Data visualization with matplotlib, seaborn, and plotly
• Feature engineering and selection techniques
• Unsupervised learning for clustering celestial objects

Each task is documented with detailed Jupyter notebooks including code, visualizations, explanations, and results.

---

📁 Repository Structure

DJS_NOVA/
├── Technical(AI)/
│   ├── Task_1/                    # Astronomical object classification & EDA
│   │   ├── task1.ipynb           # Main analysis notebook
│   │   ├── task1.md              # Task description
│   │   ├── Task 1.csv            # SDSS dataset
│   │   └── Metadata.pdf          # Dataset documentation
│   └── Task_2/                    # Stellar parameter prediction
│       ├── task2.ipynb           # Regression modeling notebook
│       ├── task2.md              # Task requirements
│       └── task2.csv             # Stellar spectra dataset
├── .gitignore                     # Git ignore rules
├── requirements.txt               # Python dependencies
├── LICENSE                        # MIT License
└── README.md                      # This file

---

✅ Completed Tasks

Task 1: Astronomical Object Classification & EDA

Dataset: SDSS (Sloan Digital Sky Survey) - 100,000 celestial objects

Objectives:

• Perform comprehensive exploratory data analysis
• Visualize celestial coordinates on 3D sphere
• Analyze photometric bands (u, g, r, i, z)
• Create color-color diagrams
• Handle missing values and outliers
• Apply unsupervised learning (clustering)

Key Findings:

• Successfully visualized 100K objects on celestial sphere
• Identified patterns in STAR, GALAXY, and QSO classes
• Removed highly correlated features (g≈u≈z, r≈i)
• Created interactive 3D visualizations with Plotly

Notebook: Technical(AI)/Task_1/task1.ipynb

---

Task 2: Stellar Parameter Prediction

Dataset: Stellar spectra with physical parameters

Objectives:

• Predict surface gravity (logg) using regression
• Predict effective temperature (Teff) using regression
• Apply feature selection techniques
• Evaluate model performance (RMSE, R², MAE)

Approach:

• Baseline: Linear Regression
• Advanced: Random Forest Regressor with hyperparameter tuning
• Feature engineering: proper motion, metallicity, photometric errors
• Outlier handling: IQR capping and percentile winsorization

Notebook: Technical(AI)/Task_2/task2.ipynb

---

🔧 Requirements

Minimum Requirements

• Python 3.8+
• pip or conda

Core Dependencies

jupyter
jupyterlab
numpy
pandas
matplotlib
seaborn
plotly
scikit-learn
scipy

Optional (for enhanced features)

umap-learn          # Better dimensionality reduction
skimpy              # Enhanced data summaries
summarytools        # Statistical summaries
polars              # Fast dataframe operations
xgboost             # Gradient boosting (bonus)

---

🚀 Quick Start

1. Clone Repository

git clone https://github.com/N1KH1LT0X1N/DJS_NOVA.git
cd DJS_NOVA

2. Create Virtual Environment

# Using venv
python -m venv .venv
.venv\Scripts\activate          # Windows
source .venv/bin/activate       # macOS/Linux

# Or using conda
conda create -n djs_nova python=3.10
conda activate djs_nova

3. Install Dependencies

pip install -r requirements.txt

4. Launch Jupyter

jupyter lab
# or
jupyter notebook

5. Open Notebooks

Navigate to:

• Technical(AI)/Task_1/task1.ipynb for Task 1
• Technical(AI)/Task_2/task2.ipynb for Task 2

---

📖 Usage

Interactive Exploration

Open notebooks in Jupyter Lab/Notebook and run cells sequentially. Each notebook is self-contained with:

• Clear section headers
• Inline comments
• Visualization outputs
• Result interpretations

Headless Execution

Run notebooks without GUI:

# Execute and save output
jupyter nbconvert --to notebook --execute Technical\(AI\)/Task_1/task1.ipynb --output task1_executed.ipynb

# Convert to HTML
jupyter nbconvert --to html Technical\(AI\)/Task_1/task1.ipynb

Key Functions & Utilities

Task 1 - EDA:

# Load and explore SDSS data
df = pd.read_csv('Task 1.csv')
from skimpy import skim
skim(df)  # Enhanced summary statistics

# 3D celestial sphere visualization
import plotly.graph_objs as go
# ... (see notebook for full code)

Task 2 - Regression:

# Standardization pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import Pipeline
from sklearn.ensemble import RandomForestRegressor

pipe = Pipeline([
    ('scaler', StandardScaler()),
    ('model', RandomForestRegressor())
])
pipe.fit(X_train, y_train)

---

📊 Results & Evaluation

Task 1 Highlights

• Data Quality: Handled 25% missing values in cam_col
• Feature Reduction: Dropped 7 highly correlated features
• Clustering: Applied KMeans, DBSCAN for unsupervised classification
• Visualizations: Interactive 3D sphere, correlation matrices, color-color diagrams

Task 2 Highlights

• Surface Gravity Model: Random Forest with optimized hyperparameters
• Temperature Model: (results pending - see notebook)
• Feature Importance: Top predictors identified
• Outlier Treatment: IQR capping with multiplier=2.5

Metrics:

• Cross-validation RMSE, R², MAE
• Test set performance
• Residual plots and diagnostics

---

🤝 Contributing

This repository represents completed recruitment tasks. Suggestions for improvements are welcome:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/improvement)
3. Commit changes (git commit -m 'Add improvement')
4. Push to branch (git push origin feature/improvement)
5. Open a Pull Request

Code Style

• Follow PEP 8 for Python code
• Use descriptive variable names
• Add comments for complex logic
• Include docstrings for functions

---

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

📧 Contact

Author: Nikhil
GitHub: @N1KH1LT0X1N
Repository: DJS_NOVA

---

🙏 Acknowledgments

• DJS Nova for providing the recruitment tasks
• SDSS for the astronomical dataset
• scikit-learn and plotly communities for excellent libraries

---

⭐ If you find this repository helpful, please consider giving it a star!