DataWhisper 📊

🔗 Live Demo: https://datawhisper-elnvwkoaq6og6w2kzwid2n.streamlit.app

A full-stack AI-powered web application built with Streamlit that allows users to upload a CSV file and automatically generates insights, visualizations, and an interactive chat interface for data exploration.

---

Features

• Exploratory Data Analysis (EDA): Automatic generation of summary statistics, missing value heatmaps, correlation matrices, and distribution/count plots.
• AI-Powered Insights: Get plain-English insights (trends, correlations, anomalies) generated by Groq AI
• Chat with Data: An interactive chat interface powered by LangChain's Pandas Dataframe Agent to ask questions directly about your data.
• Recommendations: Actionable preprocessing steps (handling missing values, outliers) suggested by AI.
• Report Export: Generate and download a comprehensive HTML report containing your visualizations and AI insights.

Screenshots

1. Login

2. Upload Data

3. Dataset Overview

4. Visual EDA

5. Missing Values Heatmap

6. Export Report

Project Structure

├── app.py                     # Main Streamlit application
├── requirements.txt           # Project dependencies
├── README.md                  # Project documentation
├── sample_data/
│   └── titanic.csv            # Sample dataset for testing
└── src/
    ├── chat.py                # LangChain Pandas Agent integration
    ├── data_loader.py         # CSV file parsing and basic info
    ├── eda.py                 # Matplotlib/Seaborn visualization logic
    ├── llm_insights.py        # LLM integration for dataset insights
    ├── recommendations.py     # LLM integration for data recommendations
    └── report_generator.py    # HTML report generation logic

Setup Instructions

Prerequisites

• Python 3.9 or higher installed (download from python.org)
• Git installed (download from git-scm.com)
• An OpenAI API key (or Groq API key) for AI features

Step 1: Clone the repository

git clone https://github.com/Payal-Dhokane/DataWhisper.git
cd DataWhisper

Step 2: Create a virtual environment (recommended)

Using a virtual environment prevents dependency conflicts with other Python projects.

On Windows:

python -m venv venv
venv\Scripts\activate

On macOS/Linux:

python3 -m venv venv
source venv/bin/activate

You should see (venv) appear in your terminal prompt, indicating the virtual environment is active.

Step 3: Install dependencies

pip install --upgrade pip
pip install -r requirements.txt

If you encounter installation errors, try installing packages one at a time:

pip install streamlit pandas numpy matplotlib seaborn plotly langchain langchain-core langchain-groq langchain-experimental tabulate PyYAML python-dotenv

Step 4: Set up API keys

Create a .env file in the project root directory:

# .env file
OPENAI_API_KEY=your_openai_api_key_here
GROQ_API_KEY=your_groq_api_key_here  # optional, alternative to OpenAI

Replace your_openai_api_key_here with your actual API key from platform.openai.com.

Step 5: Run the application

streamlit run app.py

This will open the app in your default web browser at http://localhost:8501.

Common Errors & Fixes

Error	Likely Cause	Solution
ModuleNotFoundError: No module named 'streamlit'	Dependencies not installed	Run pip install -r requirements.txt
ImportError: cannot import name '...' from 'langchain'	LangChain version mismatch	Run pip install --upgrade langchain langchain-core langchain-experimental
OpenAI API key not found	Missing .env file or API key	Create a .env file with OPENAI_API_KEY=your_key
Streamlit App - Module Not Found	Running from wrong directory	Make sure you're in the project root (cd DataWhisper)
Permission denied on venv activation (Windows)	Execution policy restriction	Run PowerShell as Administrator and execute: Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser
pip: command not found	Python/Pip not in PATH	Reinstall Python and check "Add Python to PATH" during installation

If you're still stuck, please open an issue.

Usage

1. Launch the app with streamlit run app.py
2. Upload a CSV file using the file uploader widget
3. Explore the automatically generated visualizations and statistics
4. Use the chat interface to ask questions about your data in plain English
5. Download the HTML report for sharing or documentation

Sample Data

A sample Titanic dataset is included in the sample_data/ directory to test the app without your own data.

Tech Stack

• Frontend: Streamlit
• Data Processing: Pandas, NumPy
• Visualization: Matplotlib, Seaborn, Plotly
• AI/LLM: LangChain, OpenAI / Groq
• Authentication: Streamlit-Authenticator, Streamlit-OAuth