---

Exactspace Data Science Take-Home Assignment

Author: Gunal D
Email: gunalofficialid@gmail.com
Degree: BTech Computer Science, Bangalore

---

🚀 Project Structure

Gunal_D_DataScience/
├── Task1/
│   ├── task1_analysis.py
│   ├── README.md
│   ├── shutdown_periods.csv
│   ├── anomalous_periods.csv
│   ├── clusters_summary.csv
│   ├── forecasts.csv
│   └── plots/
│        ├── data_overview.png
│        ├── shutdown_detection.png
│        ├── cluster_analysis.png
│        ├── anomaly_detection.png
│        └── forecasting_results.png
│
├── Task2/
│   ├── architecture_diagram.pptx
│   ├── notes.md
│   └── prototype/
│        ├── rag_prototype.py
│        ├── README.md
│        ├── docs/
│        └── evaluation.csv
│
├── Final_Presentation.pptx
└── CV_Gunal_D.pdf

---

📂 Task 1: Machine Data Analysis

Goals:

• Data cleaning and EDA
• Shutdown detection
• Operational state clustering
• Contextual anomaly detection
• Short-term forecasting
• Summarized actionable insights

How to Run:

1. Install dependencies:

   pip install pandas numpy matplotlib seaborn scikit-learn plotly statsmodels hdbscan pyod prophet

2. Place your data.xlsx in the Task1/ folder.
3. Run:

   python task1_analysis.py

4. Outputs:
   • Shutdown/anomaly/cluster/forecast results: .csv files
   • Key diagnostic plots: .png in plots/
   • See Task1/README.md for details.

---

📂 Task 2: RAG + LLM System Prototype

Features:

• Document chunking, semantic search (embeddings + FAISS)
• LLM-derived contextual answers with source citation
• Architecture diagram and technical notes included

How to Run:

1. Install dependencies:

   pip install sentence-transformers torch transformers faiss-cpu PyPDF2 pdfplumber

2. Add a few technical PDF docs to Task2/prototype/docs/.
3. Run:

   python rag_prototype.py

4. Outputs:
   • Interactive LLM Q&A and retrieval logs.
   • See Task2/prototype/README.md for extra usage details.

---

📑 Final Presentation

• Final_Presentation.pptx contains summary slides covering approach, workflow, findings, and results for both tasks.

---

🧑‍💻 About

This repository is the complete submission for the Exactspace Data Science Take-Home Challenge and serves as a demonstration of:

• Practical machine learning/data analysis skills
• Modern RAG/LLM workflow engineering
• Professional documentation and reproducible project structure

Feel free to clone, explore, and reach out with any questions!

---

Gunal D
gunalofficialid@gmail.com

---