3 Things: Data Analytics Highlights from July 2024

This month's '3 Things' blog post includes topics ranging from small enhancements in Power BI to examples of how the Azure OpenAI Service can impact business velocity.

Read on as we highlight three things for each of the four technology areas that you should be aware of from last month.

Power BI

• The DAX query view now enables you to add multiple measures to your Power BI semantic model, or update existing ones in one go. This can massively speed up the time it takes to create and enhance your models. Learn more.

• The web authoring experience for Power BI paginated reports has been extended to support the creation of parameters, headers, and footers including page numbers. These new capabilities are currently in preview and mean you can get a lot further than before when creating pixel-perfect reports with a compatible web browser. Learn more.

• The Enhanced row-level security editor in Power BI Desktop has been in preview for some time now and graduated to general availability status. This feature simplifies the task of adding and managing row-level security in your Power BI semantic models without the need to write any DAX. Learn more.

Microsoft Fabric

• The Microsoft Fabric team published an articled titled 'CI/CD with Warehouses in Microsoft Fabric' that explores how Fabric's Git integration and deployment pipeline features can enable continuous development and continuous deployment practices in your projects. Learn more.

• The first release of the Microsoft Fabric .NET SDK was announced. This can simplify the task of developing cross-platform applications that interact with and extend Microsoft Fabric services. It also means you can use the latest features of C# in your applications. Learn more.

• Fabric Git integration has been extended to support GitHub. This new capability is in preview and means you can choose between Azure Repos or GitHub when configuring the Git integration for your Fabric workspaces. Learn more.

Azure Analytics and AI

• Lakehouse Federation in Databricks Unity Catalog is now generally available. This capability addresses three pain points you will observe in most data estates: (1) difficulty in discovering and accessing data, (2) slow execution due to engineering bottle necks, and (3) weak compliance across siloed systems. Contributing to the general availability are three areas of improvement: (1) improved performance, (2) enhanced stability and observability, and (3) new security options. Learn more.

• There is a lot of hype around Large Language Models (LLMs). But did you know that Small Language Models (SLMs) can also provide great results? A recent episode of Microsoft's 'Data Exposed' video series takes a look at some of the advantages and reasons why you might consider them. You will also learn about Microsoft's Phi-3 models which are SLMs that provide great performance, low cost, and low latency. Learn more.

• A recent article on the Microsoft Azure AI and Machine Learning blog explores '10 ways to impact business velocity through Azure OpenAI Service'. Among the top 10, you will find a mixture of obvious use cases such as the automation of repetitive tasks and more interesting ones such as the acceleration of product development cycles. Learn more.

Open-Source Analytics

• DuckDB is exceptionally good at processing large amounts of data in memory on a single machine without running out of memory. In a recent article, DuckDB's CTO does a deep-dive into how it works while also covering some of the rationale that drove the design decisions. Learn more.

• Believe it or not, Polars has been around for four years. Since its creation, it has been widely adopted with many companies using it in production workloads today. Polars 1.0 was announced and with this the Polars team have signalled its production readiness. However, this is not the end of innovation and there are plans to significantly improve the streaming engine, GPU acceleration, introduce a 'Polars Cloud' offering, and much more. Learn more.

• Ibis is an open source dataframe library that works with data systems including DuckDB, Pandas, Polars, PySpark, and many more. A recent article on the Ibis blog explores how you can use the library to efficiently process 1TB of data on a laptop with Python. It was the pairing of Ibis with DuckDB that made it possible to run all the required queries to completion. Learn more.

Did You Find This Useful?

Get notified when we post something new by following us on X and LinkedIn.