AutoInsight AI

Multi-Agent Data Analysis Platform

AutoInsight AI is a compact, interview-ready CSV analysis app. A small LangGraph workflow coordinates specialized agents that understand the dataset, compute statistics in Python, generate practical visualizations, interpret the results with an LLM, and assemble a professional Markdown report.

The project is intentionally small so it can be explained confidently in 15 to 20 minutes.

Project Overview

Users upload a CSV file in Streamlit, then the workflow runs end to end:

Upload CSV -> Supervisor -> Data Understanding -> Statistics -> Visualization -> Insight -> Report -> End

Architecture Diagram

flowchart LR
    UI[Streamlit UI] --> G[LangGraph Workflow]
    G --> S[Supervisor Agent]
    S --> D[Data Understanding Agent]
    D --> T[Statistics Agent]
    T --> V[Visualization Agent]
    V --> I[Insight Agent]
    I --> R[Report Agent]
    R --> O[Markdown Report]

Loading

Workflow Diagram

flowchart TD
    A[Upload CSV] --> B[Supervisor]
    B --> C[Data Understanding]
    C --> D[Statistics]
    D --> E[Visualization]
    E --> F[Insight]
    F --> G[Report]
    G --> H[End]

Loading

Tech Stack

• Python
• LangGraph
• LangChain
• Groq
• Pandas
• Plotly
• Matplotlib
• Pydantic
• Streamlit

Project Structure

AutoInsight-AI/
├── app/
│   ├── agents/
│   │   ├── supervisor.py
│   │   ├── data_agent.py
│   │   ├── statistics_agent.py
│   │   ├── visualization_agent.py
│   │   ├── insight_agent.py
│   │   └── report_agent.py
│   ├── tools/
│   │   ├── dataframe.py
│   │   ├── statistics.py
│   │   └── plotting.py
│   ├── graph.py
│   ├── state.py
│   └── llm.py
├── frontend/
│   └── streamlit_app.py
├── report/
├── requirements.txt
└── README.md

Installation

1. Create and activate a virtual environment.
2. Install the dependencies:

pip install -r requirements.txt

3. Set your Groq API key:

set GROQ_API_KEY=your_key_here

On PowerShell:

$env:GROQ_API_KEY = "your_key_here"

How to Run

Start the Streamlit app from the project root:

streamlit run frontend/streamlit_app.py

Upload a CSV file, click Run Analysis, and review the overview, statistics, visualizations, insights, and downloadable Markdown report.

Screenshots

Dashboard

Statistics

Insight 1

Insight 2

Insight 3

Visualization 1

Visualization 2

Visualization 3

Generated Report

Sample Generated Report

The AI-generated report can be downloaded directly from the application. A sample generated report is available here:

📄 View Sample Markdown Report

---

Notes

• Statistics are computed in Python only.
• The LLM is used only for interpretation.
• Invalid plots are skipped instead of crashing the workflow.

Future Improvements

• Add optional caching for repeated analyses.
• Add support for XLSX files.
• Add time-series specific visualizations when datetime columns are present.
• Save rendered reports automatically into the reports/ folder.
• Add test coverage for the data, statistics, and plotting tools.

---

Author

Apurva Mishra
IMSc Quantitative Economics & Data Science
Birla Institute of Technology, Mesra

GitHub: https://github.com/apooorv19
LinkedIn: https://www.linkedin.com/in/apooorv/

---

Credits

Iris Dataset

Fisher, R. A. (1936).
The use of multiple measurements in taxonomic problems.
Annals of Eugenics, 7(2), 179–188.

Dataset available from the UCI Machine Learning Repository:
https://archive.ics.uci.edu/ml/datasets/iris