Skip to content

aidatacooper (DataCooper AI Lab)

Welcome to aidatacooper — the open-source engineering lab founded by Cooper Wenhua.

Our mission is to make Business Intelligence programmable, agent-friendly, and reproducible. We bridge the gap between LLMs, Model Context Protocol (MCP), and Tableau to transition BI workflows from manual, error-prone drag-and-drop tasks into inspectable and safe code-first systems.


The datacooper Story: Why Programmable BI?

For years, Business Intelligence has been locked behind graphical user interfaces (GUIs). While great for visual exploration, this drag-and-drop paradigm introduces major challenges for modern software teams:

  • No Version Control: .twb and .tfl files are XML/JSON monoliths that are difficult to diff, track, or merge in Git.
  • Agent Barriers: AI agents cannot easily click through desktop menus to build charts or map calculations.
  • Lack of Validation: There are few automated ways to ensure a workbook calculation doesn't break when a schema changes.

datacooper was born to solve this. We build PyPI packages, CLI compilers, and MCP servers that treat Tableau workbooks and Prep flows as reproducible analytics assets.


Our Open-Source Ecosystem

We build and maintain the core engines that turn manual BI steps into programmable systems:

Featured Repositories

Project Target The Core Innovation
cwtwb Tableau Workbook (.twb / .twbx) Text-to-Workbook compiler and parser. Allows AI agents (via MCP) to dynamically edit worksheets, manage calculated fields, and validate schemas without GUI interaction.
cwprep Tableau Prep Flow (.tfl) Text-to-PrepFlow engine. Translates natural language data-cleaning requests into valid Tableau Prep flows, offering offline validation, schema checking, and automated ETL staging.

🔌 The MCP Paradigm (Model Context Protocol)

We believe AI agents should have native tools to inspect and edit your BI layer. Both cwtwb and cwprep expose standardized Model Context Protocol (MCP) endpoints, turning local Tableau files into contexts that LLMs (like Claude and OpenAI apps) can reason about, refactor, and build.

// Hook datacooper tools directly into Claude Desktop or Cursor:
{
  "mcpServers": {
    "cwtwb": {
      "command": "uvx",
      "args": ["cwtwb"]
    },
    "cwprep": {
      "command": "uvx",
      "args": ["cwprep"]
    }
  }
}

🔮 The Stack We Stand On


💬 Connect & Collaborate

Popular repositories Loading

  1. cwtwb cwtwb Public

    A Python-based engine that enables Text-to-Tableau twb dashboards generation.

    Python 42 2

  2. cwprep cwprep Public

    A Python-based engine that enables Text-to-Tableau prep generation.

    Python 5 3

  3. tableau-monitor tableau-monitor Public

    Auto monitor and cancel long-running Tableau Server flows

    Python

  4. 20260714-tableaudesign 20260714-tableaudesign Public

    HTML

  5. .github .github Public

  6. wow-tableau-problem-solving wow-tableau-problem-solving Public

    AI-assisted Tableau Workout Wednesday case analysis and cwtwb replication lab

    Python

Repositories

Showing 6 of 6 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…