2 Hours (Hands-On Lab Format) Power BI Developers familiar with semantic models, new to MCP VS Code + GitHub Copilot + Power BI Desktop Sales & Returns Sample v201912.pbix (download from Microsoft)
概览
2 Hours (Hands-On Lab Format) Power BI Developers familiar with semantic models, new to MCP VS Code + GitHub Copilot + Power BI Desktop Sales & Returns Sample v201912.pbix (download from Microsoft) - 2 Hours | Hands-On Lab | VS Code + Copilot - AI-Powered Semantic Model Development - github.com/microsoft/powerbi-modeling-mcp - What is MCP? Protocol fundamentals - Power BI Modeling MCP architecture - Install VS Code extension - Configure GitHub Copilot - Download sample report & save as PBIP - Set up two-folder workspace - Connect MCP to Power BI Desktop Everyone connected and making their first MCP call - Bulk rename columns & measures - Add descriptions across model - Create KPI Measures table with complex DAX - Build Product Performance calculated table - Add fiscal calendar columns & time intelligence - Create Calculation Groups (Time & Currency) - Implement Row-Level Security (static roles by store type/product category) - Create Perspectives & generate documentation - Git...
README
以下为原文 README 节选，完整内容请查看原文。
Power BI Modeling MCP Server Workshop
Deep Dive: AI-Powered Semantic Model Development
Duration: 2 Hours (Hands-On Lab Format)
Audience: Power BI Developers familiar with semantic models, new to MCP
Tools: VS Code + GitHub Copilot + Power BI Desktop
Sample Report: Sales & Returns Sample v201912.pbix (download from Microsoft)
HOUR 1: Setup & Architecture (~60 minutes)
Title
Power BI Modeling MCP Server
Deep Dive Workshop
- 2 Hours | Hands-On Lab | VS Code + Copilot
- AI-Powered Semantic Model Development
- github.com/microsoft/powerbi-modeling-mcp
Workshop Agenda
Hour 1 — Setup & Architecture
- What is MCP? Protocol fundamentals
- Power BI Modeling MCP architecture
- Install VS Code extension
- Configure GitHub Copilot
- Download sample report & save as PBIP
- Set up two-folder workspace
- Connect MCP to Power BI Desktop
Goal: Everyone connected and making their first MCP call
Hour 2 — Hands-On Lab
- Lab 1: Bulk rename columns & measures
- Lab 2: Add descriptions across model
- Lab 3: Create KPI Measures table with complex DAX
- Lab 4: Build Product Performance calculated table
- Lab 5: Add fiscal calendar columns & time intelligence
- Lab 6: Create Calculation Groups (Time & Currency)
- Lab 7: Implement Row-Level Security (static roles by store type/product category)
- Lab 8: Create Perspectives & generate documentation
- Lab 9: Git commit, push, and rollback
- Lab 10: Fix report visual references (separate VS Code window)
Goal: Complete real-world model changes with CI/CD safety net
What is MCP?
Model Context Protocol — The bridge between AI and your tools
MCP is an open protocol that standardizes how AI assistants connect to external tools, data sources, and services. Think of it as a universal adapter between LLMs and the real world.
Three Core Components:
| Component | Description |
|---|---|
| TOOLS | Functions the AI can call to perform actions (create measure, modify table, execute DAX) |
| RESOURCES | Data and context the AI can read (model metadata, DAX patterns, documentation) |
| PROMPTS | Pre-built templates for common tasks (connect, query, analyze performance) |
Key Insight
MCP servers run locally on your machine. Your data never leaves your environment — the AI sends commands, not your data.
Reference: modelcontextprotocol.io
Architecture Overview
Data Flow Diagram
┌─────────────────────────────────┐
│ YOU │
│ VS Code + GitHub Copilot │
│ (Natural language prompts) │
└───────────────┬─────────────────┘
│
│ MCP Protocol (stdio)
▼
┌─────────────────────────────────┐
│ LOCAL SERVER │安装
This server does not publish a one-line install command.
Open the repository installation guide 新手？先看 mcpskill.org 教程 →