20200915 - Power BI AI Features - Meetup

Transcript

1. Power BI AI Visuals & Capabilities Microsoft AI & ML Community Tuesday, 15 Sept 2020
2. About Aroh Technical Consultant, Microsoft MVP, MCT & Nintex vTE Experience: Office 365, SharePoint Online, Data Platform Office 365 Consultant & Power Platform Trainer LinkedIn: https://www.linkedin.com/in/arohshukla/ Blog: aarohblah.blogspot.com @aaroh_bits Aroh Shukla
3. Agenda • Overview of AI in Power BI • End User Insights • AI Visuals • Overview of Key Influencers Visual • Overview of Decomposition Tree Visual • Overview of Q&A Visual • Demos: Key Influencers Visual, Decomposition Tree Visual, Q&A Visual • Q and A • (Slide Deck, Demo Samples and Blog will be shared after Event )
4. Investment areas of AI in Power BI Sentiment Analysis Key Phrase Extraction Create ML models Explore predictions Python Integration R Integration Extend with Azure ML Integrate into reports AI Visualizations Natural Language
5. Power BI is a business analytics service that delivers insights to enable fast, informed decisions with stunning visuals. Any data, any time, anywhere Whatis Power BI? 1. Power BI Desktop 2. Power BI Service 3. Power BI Mobile
6. Microsoft Power BI Desktop
7. What is Dataflow? Many definitions about Power BI Dataflow • Self-service data prep or Self-service ETL (Extract-Transform-Load) • Way that is reusable & repeatable for others • A collection of entities • Entities are similar to tables. • Definition • Dataflow is simply Power Query in the cloud. • Power Query process that runs in cloud independently from any Power BI reports. • Dataflows are used to ingest, transform, integrate, and enrich big data/AI • A lot of ML capabilities are built on top of Dataflow. (More computation power on cloud than Desktop)
8. What is Dataflow?
9. Power BI Pro vs Premium
10. 1st Key Influencers Visual When to use key influencers • Factors affect the metric being analysed. • Contrast the relative importance of these factors. How AI helps • Visual a couple of algorithm like ML.NET, One-hot encoding, Replace missing value etc.
11. Key Influencers Visual – Limitations Considerations • Direct Query is not supported • Azure Analysis Services and SSAS is not supported • Publish to Web is not supported • .NET FX 4.6 above or higher is required • Aggregates and measures are not supported.
12. 2nd Decomposition Tree Visual Visualize data • across multiple dimensions • automatically aggregates data • ad hoc exploration and conducting root cause analysis. Visual Inputs • Analyse – metric to analyse • Explain By: one or more dimensions you would like to drill down into. AI splits • High Value: highest value of the measure being analysed. • Low Value: lowest value of the measure being analysed.
13. Decomposition Tree Visual – Limitations The decomposition tree is not supported • On-premises Analysis Services AI splits are not supported • Azure Analysis Services • Direct Query • Power BI Report Server • Publish to Web • Complex measure
14. 3rd Q&A Visual Visualize data • allows users to ask natural language questions and get answers in the form of a visual. • both Power BI Desktop and the Power BI service. Q&A Visual 4 core complements • Question box. • Pre-populated list of suggested questions. • Icon to convert the Q&A visual into a standard visual. • Icon to open Q&A tooling which allows designers to configure the underlying natural language engine.
16. 3rd Q&A Visual - Limitations Data sources not supported • Object level security with any type of data source. • DirectQuery against any source • Composite models (DirectQuery + Import data) • Reporting Services Other Considerations • Tooling limitations • Review question limitations • Teach Q&A limitations • Refer to this link to get in-depth about Q&A visual.