AI-DataScience-Lab

AI-DataScience-Lab is an end-to-end forecasting web application designed to upload CSV datasets, clean and analyze them using Python libraries, generate visualizations and predictive models with scikit-learn, and summarize the dataset using OpenAI’s GPT-3.5 API.

The frontend is hosted on GitHub Pages, and the backend is deployed on Azure App Service, creating a scalable and professional architecture suitable for real-world AI and data science workflows.

---

🌐 Live Demo

• 
• Frontend (GitHub Pages): https://hariprashad-ravikumar.github.io/AI-DataScience-Lab/
• Backend (Azure)

---

⚙️ Features

• Upload CSV files with two columns: X (dates) and Y (numerical values)
• Cleans data using pandas, removes invalid entries
• Generates a scatter plot using matplotlib
• Converts date strings to ordinal format and trains a LinearRegression model with scikit-learn
• Uses OpenAI API (GPT-3.5-turbo) to summarize the uploaded dataset
• Predicts future Y values for user-supplied future X (date) values
• Secure HTTPS communication across GitHub and Azure (CORS-enabled)
• Temporary file storage using Python's tempfile, cleaned automatically on restart

---

📊 Technical Workflow

1. Frontend (GitHub Pages)

• HTML + JavaScript app with forms to:
  • Upload CSV data
  • Request future predictions
• Communicates with the backend via fetch() using HTTPS POST requests
• Displays:
  • Processing log
  • OpenAI-generated summary
  • Forecast output
  • Auto-generated plot image

2. Backend (Azure App Service - Python Flask)

• Routes:
  • POST /upload: Handles file uploads, data cleaning, modeling, summary generation
  • POST /predict: Accepts future dates, returns predictions
  • GET /plot.png: Serves saved scatter plot image

3. Processing Pipeline

• Step 1: Data Cleaning

  • Reads CSV using pandas
  • Drops NA values and converts X to datetime format

• Step 2: Visualization

  • Uses matplotlib to generate scatter plot
  • Plot saved to a temporary directory and served on request

• Step 3: Modeling

  • Uses scikit-learn LinearRegression to fit X (date ordinal) → Y
  • Model used to predict future values based on user input

• Step 4: Summarization

  • Sends cleaned dataset (via .head(10).to_csv()) to OpenAI GPT-3.5 API
  • Summary generated and returned to frontend

---

🛠️ Tech Stack

Layer	Technology
Frontend	HTML, JavaScript, GitHub Pages
Backend	Flask, Azure App Service
ML Tools	pandas, scikit-learn, matplotlib
AI	OpenAI GPT-3.5 (openai Python SDK)
Storage	Python tempfile for secure cleanup
Deployment	Gunicorn + Azure Linux App Container

---

🔐 Security and Performance

• Uses flask-cors to securely allow cross-origin requests from GitHub Pages
• All requests are served over HTTPS
• Files and plots are saved temporarily and deleted automatically on app shutdown using tempfile.TemporaryDirectory and atexit

---

🚀 How to Run Locally

1. Clone the repo:

   git clone https://github.com/Hariprashad-Ravikumar/AI-DataScience-Lab.git
   cd AI-DataScience-Lab/backend