Open-Source DATAGEN Uses 8 Specialized Agents to Automate Data Science Reports
An open-source multi-agent stack orchestrates hypothesis generation, code execution, visualization, and report writing under a LangGraph state machine with human review.
PRO- DATAGEN is an open-source multi-agent research assistant with 1.7k GitHub stars, MIT licensed on GitHub.
- Eight specialized agents cover hypothesis, code, visualization, search, reporting, quality review, and note-taking.
- LangGraph state machine orchestrates the workflow with a human review gate before heavy processing.
- Each agent can be pinned to a different provider: OpenAI, Anthropic, Google, Ollama, or Groq.
- Progressive Disclosure architecture and MCP integration bring filesystem, GitHub, and Tavily web search.
- Known weak spots include NoteTaker efficiency, total runtime, and potential in-place modification of input data.
A new open-source project called DATAGEN is picking up traction as a multi-agent research assistant that tries to automate the messy middle of data science work: forming hypotheses, cleaning data, writing analysis code, generating plots, searching literature, and drafting the report. It bills itself as an AI-powered platform that uses specialized agents for analysis, visualization, and reporting, built on LangChain, OpenAI GPT models, and LangGraph.
The repo has crossed 1.7k stars, ships under an MIT license, and is written entirely in Python. It was previously known as AI-Data-Analysis-MultiAgent before being rebranded.
A crew of agents, not a monolith
Rather than running a single prompt-and-pray loop, DATAGEN splits the workflow across eight named agents, each with a scoped role:
hypothesis_agentgenerates research hypothesesprocess_agentsupervises the entire research processvisualization_agentcreates data visualizationscode_agentwrites data analysis codesearcher_agentconducts literature and web searchesreport_agentwrites research reportsquality_review_agent
This story is for Pro members
You've reached the end of the free preview. Upgrade to AlphaSignal Pro to read the full article - and everything else behind the paywall.