International IT Solutions B.V. logo
IITS.nlAI & Data
All training courses
Advanced2 days

IoT Analytics on Azure: From Devices to Dashboards

End-to-end device telemetry pipelines with IoT Hub, Azure Data Explorer, and Digital Twins

Overview

Connected devices generate the most time-sensitive, highest-volume data in the enterprise — and most data teams are not equipped to handle it. This two-day training teaches the Azure IoT analytics reference architecture published in the Azure Architecture Center: IoT Hub and Event Hubs for device ingestion, Azure Functions and Stream Analytics for real-time routing and enrichment, Cosmos DB for the operational hot path, and Azure Data Explorer (ADX) for the analytical cold path where KQL unlocks anomaly detection, time-series forecasting, and predictive maintenance at petabyte scale. You will also model physical assets with Azure Digital Twins and build Grafana and Power BI dashboards that give operators a live, unified view of their physical world.

What you'll learn

  • Design a dual-path IoT analytics architecture with a hot path (operational, low-latency) and a cold path (analytical, high-throughput) using Azure managed services
  • Ingest device telemetry from IoT Hub and Kafka into Azure at scale, handling device identity management, D2C messaging, and message routing
  • Process and enrich event streams in real time using Azure Functions and Azure Stream Analytics before splitting to Cosmos DB and ADX
  • Query, visualise, and act on device telemetry in Azure Data Explorer using KQL — including anomaly detection, time-series analysis, and forecasting functions
  • Model physical assets, spaces, and their relationships as a live graph using Azure Digital Twins, then stream twin state changes into ADX for analytics
  • Build operational monitoring dashboards with ADX native dashboards, Grafana, Power BI, and Jupyter notebooks connected to ADX

Programme

Day 1 — Device ingestion, routing & the operational path
  • Azure IoT analytics reference architecture: hot path vs cold path design decisions and when each matters
  • Azure IoT Hub: device provisioning service, D2C and C2D messaging, device twins, and message routing to multiple endpoints
  • Event Hubs as an IoT ingestion layer: Kafka protocol support, consumer groups, capture to ADLS Gen2, and scaling for millions of events per second
  • Azure Functions for real-time event processing: event-triggered functions, binding to Cosmos DB output, and fan-out patterns for multi-destination routing
  • Azure Stream Analytics: windowing functions (tumbling, hopping, sliding), reference data joins for device metadata enrichment, and anomaly detection operators
  • Cosmos DB as the operational hot path: NoSQL document model for device state, TTL for automatic data expiry, and change feed integration with Functions
  • Hands-on: build a complete IoT ingestion pipeline from IoT Hub through Stream Analytics to Cosmos DB, with a Functions-triggered alert on threshold breach
Day 2 — Azure Data Explorer, Digital Twins & dashboards
  • Azure Data Explorer architecture for IoT: tables, ingestion policies, data partitioning, and continuous ingestion from Event Hubs using native connectors
  • KQL for IoT analytics: time-series operators (make-series, series_decompose), anomaly detection with series_decompose_anomalies, and multi-step forecasting
  • Predictive maintenance patterns in ADX: rolling window statistics, threshold alerting using update policies, and feeding KQL results into Azure Machine Learning for scoring
  • Azure Digital Twins: modelling physical assets with DTDL, creating twin graphs for buildings, factories, or vehicle fleets, and streaming twin state changes to Event Hubs and ADX
  • Visualisation layer: ADX native dashboards for operations teams, Grafana integration with the ADX data source plugin, Power BI for business reporting, and Jupyter notebooks for ad-hoc analysis
  • Security and governance: managed identity authentication between services, private endpoints for IoT Hub and ADX, Microsoft Defender for IoT for OT/IT network monitoring
  • Production architecture: high-availability patterns, disaster recovery for ADX clusters, cost management with ADX auto-stop and tiered caching policies
  • Hands-on: ingest simulated IoT telemetry into ADX, run KQL anomaly detection queries, model twin relationships in Azure Digital Twins, and build a Grafana dashboard for real-time device monitoring

Who is this for?

  • Data engineers designing cloud-side analytics pipelines for IoT or industrial telemetry data
  • IoT solution architects evaluating Azure services for smart building, fleet management, or manufacturing analytics
  • Analytics engineers extending existing Azure data platforms to handle device-generated time-series data
  • Platform teams in energy, utilities, logistics, or healthcare managing large fleets of connected assets

Prerequisites

  • Azure data services experience (Event Hubs, Storage, or similar)
  • Python or C# programming fundamentals
  • Basic understanding of IoT concepts: devices, telemetry, D2C messaging, and MQTT/AMQP protocols

Tools & technologies covered

Azure IoT HubAzure Event HubsAzure Data ExplorerAzure Stream AnalyticsAzure FunctionsAzure Cosmos DBAzure Digital TwinsKQLGrafanaPower BIAzure Machine Learning
Not sure which course fits your team?
Talk to us — we'll match you to the right training path.
Get in touch