We'll continue to delve into the Azure Data Architecture Guide with our fifth blog entry in this series. The previous entries for this blog series are:

• Azure Data Architecture Guide – Blog #1: Introduction
• Azure Data Architecture Guide – Blog #2: On-demand big data analytics
• Azure Data Architecture Guide – Blog #3: Advanced analytics and deep learning
• Azure Data Architecture Guide – Blog #4: Hybrid data architecture

Like the previous post, we'll work from a technology implementation seen directly in our customer engagements. The example can help lead you to the ADAG content to make the right technology choices for your business.

Clickstream analysis

Engage with your customers and uncover insights from data generated by clickstream logs in real-time, using Azure. Rapidly ingest incoming data through Event Hubs (or leverage Apache Kafka), process it with Spark streaming and Spark ML for predicting product recommendations, then use the Spark to Azure Cosmos DB connector to save the processed data to Cosmos DB for global distribution to your customers, wherever they are.

As an alternative to Apache Sparks on HDInsight (Spark ML), you can use Azure Databricks—a fast, easy, and collaborative Apache Spark–based analytics service.

Highlighted services

• Azure Event Hubs
• HDInsight with Apache Spark
• HDInsight with Kafka
• Azure Cosmos DB

Related ADAG articles

• Big data architectures
  • Big data overview
  • Real time processing
• Scenarios
  • Working with CSV and JSON files for data solutions
  • Processing free-form text for search
  • Natural language processing
• Technology choices
  • Choosing a search data store in Azure
  • Choosing a real-time message ingestion technology in Azure
  • Choosing a natural language processing technology in Azure

Please peruse ADAG to find a clear path for you to architect your data solution on Azure:

• Read the Azure Data Architecture Guide

Azure CAT Guidance

"Hands-on solutions, with our heads in the Cloud!"