🤖 AI EDA Assistant

A smart Exploratory Data Analysis (EDA) assistant that automates data profiling, visualization, and summarization using Python and AI-powered tools.

---

🚀 Project Overview

This project is designed to assist data scientists and analysts by automating the most common EDA tasks. With just a few lines of code, the assistant can generate data summaries, detect nulls, display feature distributions, correlations, and offer insights using natural language.

---

💻 Tech Stack

• Python
• Pandas / NumPy – For data manipulation and statistical calculations
• Matplotlib / Seaborn – For plotting and visualizations
• Streamlit – To build an interactive web interface
• YData Profiling – For automatic EDA reports
• scikit-learn – For preprocessing and basic ML utilities

---

📄 What's Inside

• app.py
  The Streamlit application that powers the assistant.

• eda_tools.py
  Custom Python script for handling EDA logic and feature generation.

• requirements.txt
  Python dependencies required to run the application:

  streamlit
  pandas
  numpy
  matplotlib
  seaborn
  ydata-profiling
  scikit-learn
  

• sample_datasets/
  Includes example datasets you can use to test the tool.

---

🧠 Features

• Upload CSV files and get:
  • Dataset overview (shape, dtypes, missing values)
  • Descriptive statistics
  • Correlation matrix and heatmap
  • Class balance checks (for classification tasks)
  • Automated report using YData Profiling
• Simple, interactive controls using Streamlit sidebar

---

📦 How to Run Locally

1. Clone the repository:

   git clone https://github.com/sjapanjots/AI_EDA_Assistant.git
   cd AI_EDA_Assistant

2. Install dependencies:

   pip install -r requirements.txt

3. Run the Streamlit app:

   streamlit run app.py

4. Go to http://localhost:8501 in your browser.

---

🖼️ Sample Use Case

• Upload: iris.csv
• Output:
  • Data summary
  • Target distribution
  • Feature correlation heatmap
  • Automated profiling report (HTML)

---

🙋‍♂️ Author

Japanjot Singh
Data Scientist & ML Enthusiast
📬 sjapanjots@gmail.com