data_source_instructions.md

Data Source Instructions

Overview

This Fabric Data Agent has access to manufacturing operations data in an EventHouse database.

Business Context

Note: Replace this content with your organization's actual business performance measurements.

This sample dataset represents common manufacturing scenarios:

• Equipment monitoring and maintenance
• Product quality control
• Operational efficiency analysis
• Real-time alerting and diagnostics

Database Tables and Relationships

Note: Replace this content with your organization's actual data and update table schemas accordingly.

Core Tables

• events - Manufacturing telemetry and sensor data (real-time + historical)
• assets - Equipment and machinery information
• products - Product catalog and specifications
• sites - Manufacturing facility locations
• locations - Geographic information

Data Relationships

events → assets → sites → locations
events → products

Real-Time Data

• events table receives continuous real-time data via EventStream
• Contains sensor readings: temperature, vibration, humidity, speed
• Includes quality metrics: defect probability
• Links to specific assets and products

Reference Data

• Static tables (assets, products, sites, locations) contain stable reference information
• Used for context and enrichment of event data
• Updated infrequently

Query Patterns

When analyzing manufacturing data, typically join events with reference tables:

// Asset performance analysis
events
| join assets on $left.AssetId == $right.Id
| summarize avg(Temperature), avg(Speed) by AssetId, assets.Name

// Product quality tracking  
events
| join products on $left.ProductId == $right.Id
| summarize avg(DefectProbability) by ProductId, products.Name