Repolode
Submit
zi-yue-1129

zi-yue-1129/DATAGEN

Multi-agent research assistant built on LangGraph that generates hypotheses, runs data analysis and visualization, and writes research reports.

View on GitHub
Stars
1,814
Forks
257
Open issues
8
Last commit
4 days ago

What is it?

DATAGEN is a Python platform that uses several specialized LLM agents to carry out a data research workflow. Given a data file and a research question, it proposes hypotheses, lets the user choose whether to continue or regenerate them, then coordinates agents that write analysis code, produce charts, search the web or literature, and draft a report. A quality review agent checks the output and triggers revisions when needed.

The workflow is modeled as a LangGraph state graph, and a dedicated note agent records the research process as it runs. Each agent's model provider and parameters are set in an agent_models.yaml file, with support for OpenAI, Anthropic, Google, Ollama, Groq and other providers. Tools and skills are configured through a config directory, and external capabilities such as filesystem, GitHub and web search are connected through the Model Context Protocol (MCP).

The project needs Python 3.10 or higher and API keys for the model providers you use. The README warns that agents may modify the analyzed data, so backing up input files first is recommended.

Key features

  • 01Generates and refines research hypotheses with a human choice step.
  • 02Writes data analysis code and produces visualizations.
  • 03Searches the web and literature, then drafts research reports.
  • 04Runs a quality review and revision loop on the output.
  • 05Lets each agent use a different LLM provider, configured in YAML.
  • 06Connects external tools through the Model Context Protocol (MCP).

Use cases

  • Running exploratory analysis on a CSV dataset and getting a written report with charts.
  • Prototyping an automated research pipeline built on LangGraph and multiple LLM providers.
  • Studying how a multi-agent system splits up a data analysis task.

Quick start

  1. 1

    Clone the repository

    git clone https://github.com/starpig1129/DATAGEN.git
  2. 2

    Create and activate a Conda virtual environment

    conda create -n datagen python=3.10 && conda activate datagen
  3. 3

    Install dependencies

    pip install -r requirements.txt
  4. 4

    Rename .env Example to .env, fill in the values, then run main.py

    python main.py

Things to know

  • Requires Python 3.10 or higher and a Conda environment setup.
  • Needs API keys for language model providers and ChromeDriver configured in environment variables.
  • The agent system may modify the data being analyzed, so backing up data first is recommended.
  • Runs make multiple API calls, so sufficient API credits are needed.

In collections

Open-source AI agent frameworks & appsData science courses & tools

Topics

#agent#multi-agent#data-analysis#langgraph#llm#langchain#data-science#code-generation#ai#ai-data-analysis

FAQ

What is DATAGEN?
Multi-agent research assistant built on LangGraph that generates hypotheses, runs data analysis and visualization, and writes research reports. DATAGEN is a Python platform that uses several specialized LLM agents to carry out a data research workflow. Given a data file and a research question, it proposes hypotheses, lets the user choose whether to continue or regenerate them, then coordinates agents that write analysis code, produce charts, search the web or literature, and draft a report. A quality review agent checks the output and triggers revisions when needed.
Is DATAGEN free to use? What license does it have?
DATAGEN is open source under the MIT license. Check the license terms before using it in commercial products.
Is DATAGEN still maintained?
Actively maintained. The last commit to DATAGEN was on Nov 2, 2026.
What can I use DATAGEN for?
Running exploratory analysis on a CSV dataset and getting a written report with charts. Prototyping an automated research pipeline built on LangGraph and multiple LLM providers. Studying how a multi-agent system splits up a data analysis task.
How do I install DATAGEN?
Clone the repository: git clone https://github.com/starpig1129/DATAGEN.git Create and activate a Conda virtual environment: conda create -n datagen python=3.10 && conda activate datagen Install dependencies: pip install -r requirements.txt Rename .env Example to .env, fill in the values, then run main.py: python main.py
Where can I get DATAGEN?
The source code and installation instructions are on GitHub: https://github.com/zi-yue-1129/DATAGEN

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects