You are on page 1of 26

Cristian Andrei Lefter

http://about.me/CristianLefter

AI in Power BI
Sponsors
GOLD Sponsor
Raffle Sponsors

Print Sponsor

Organized by
Agenda
• Introduction
• The Unified BI Platform
Introduction
Artificial Intelligence Terms
Artificial Intelligence
covers anything which enables
Artificial Intelligence computers to behave like
humans.

Machine Learning is the


subset of AI which uses
Machine Learning statistical methods to enable
machines to improve with
experiences.

Deep Learning is a specific class


of Machine Learning algorithms
fx which use complex neural
networks. It allows software to train
Deep Learning
itself to perform tasks.
Power BI history

Project Gemini Power Pivot Project Crescent Power View Power Query PowerBI.com Premium
In Memory Excel Add-in Self-Service BI Part of Share Power BI for Power BI Embedded
Engine capabilities in Point Office 365 Service Power BI
Excel Report Server

2006 2009 2010 2012 2013 2015 2017


Power BI portfolio
Unified BI Platform
Unified BI Platform
Unified BI Platform

Self Service Data Prep Pervasive


for Big Data Artificial Intelligence for BI
Unifying Self-Service & Enterprise BI
Self-Service BI Models

Interactive reports & dashboards

Enterprise grade semantic models

Enterprise reporting
Unified BI Platform

Unified platform for both Pervasive


Self-Service and Enterprise BI Artificial Intelligence for BI
What are dataflows?
Power BI Dataflows
- Were previously called Common Data Service for Analytics as well as
Datapools.
- Dataflows are a self-service ETL tool in the Power BI cloud service
- Not Azure Data Factory
- Not Microsoft Flow
- Not a Dataset
- Effectively: Power Query in the cloud/in the browser plus a lot more
- They split the extraction and preparation of data away from Power BI
datasets
- They allow for sharing of tables of data between datasets
- They do not allow sharing of dataset features like relationships and
DAX calculations
- A collection of entities
- Entities are similar to tables
Why would I use a Power BI Dataflow?
Easily ingest, cleanse, transform, integrate, enrich, and schematize data.

Traditional ETL (SSIS)

Power BI
Dataflows 014
Unified BI Platform

Unified platform for both Integrating the broader analytics


Self-Service and Enterprise BI platform with Power BI
AI in Power BI – the short story

CAPABILITIES CAPABILITIES CAPABILITIES LANGUAGES INTEGRATION

• Quick Insights • Sentiment analysis • Predictions, • R • Azure ML


• Q&A • Key phrase extraction Classifications,
• Python • Azure Frameworks
• AI Visualizations • OCR, Text Translations • Forecasting, Clustering,
• Recommendations
Natural Language exploration +Automated Insights

Demo
Python and Power BI

Demo
Automated Machine Learning (AutoML) for
Dataflows in Power BI
AutoML enables business analysts to build machine learning models with clicks, not code, using just their Power BI
skills. With AutoML, the data science behind the creation of ML models is automated by Power BI, with guardrails to
ensure model quality, and visibility to ensure you have full insight into the steps used to create your ML model.
Automated ML notes
Automated ML in Power BI Automated ML in Azure
Designed for the business analyst/
Designed for for the data scientist/data engineer
BI professional
Skillset required is Azure, Python
Skillset required is Power Query
and Jupyter notebooks
and Power BI dataflows (no code)

ML model lives in Azure and can be consumed


ML model lives completely inside Power
as a webservice by lots of different applications
BI and can only be used by Power BI
(including Power BI)
Requires Power BI Premium Requires an Azure subscription and utilizes
and utilizes capacity Azure consumption
Cognitive Services in Power BI
Power BI provides access to a set of functions from Azure Cognitive Services to enrich your data in the self-service data
prep for Dataflows. The services that are supported today are Sentiment Analysis, Key Phrase Extraction, Language
Detection, and Image Tagging. The transformations are executed on the Power BI service and do not require an Azure
Cognitive Services subscription. This feature requires Power BI Premium.
Cognitive services notes
Cognitive services in Power BI Cognitive services in Azure
Designed for the business analyst/
Designed for for the data scientist/data engineer
BI professional
Skillset required is Power Query Skillset required is Azure, coding language(C#,
and Power BI dataflows (OOB point and Java, Node.js Python)
click functions, M code)
Lives completely inside Power BI and Lives in Azure and can be consumed as a
can only be used by Power BI webservice by lots of different applications
Requires Power BI Premium Requires an Azure subscription and utilizes
and utilizes capacity Azure consumption
Data + AI on Azure Data Platform
Dynamics Power BI Azure Azure Azure Azure
365 Data Databricks SQL DW ML
Factory
AI consumption Self-service data prep Data ingestion Enterprise Curated data AI models
AI for Sales Dataflows & orchestration data prep Auto ML

AI consumption

Enterprise BI
Semantic models

Self-service BI

Contact Lead Opportunity Account Customer Profile Product Profile People Profile
Session Resources
• Power BI blog: https://aka.ms/PBIBlog
• Testing Power BI Premium Features With Power BI Embedded
https://blog.crossjoin.co.uk/2018/10/14/testing-power-bi-premium-
features-with-power-bi-embedded/
• Trillion row demo: https://aka.ms/TrillionRowDemo
CristianLefter@hotmail.com
http://about.me/CristianLefter
Ne vedem la anul!
See you next year!
• Reserved date: 6th of June 2020
Or this year in Bucharest
• 5th October 2019
https://www.sqlsaturday.com/915/eventhome.aspx

Please evaluate THIS event!


• Overall event https://www.sqlsaturday.com/874/EventEval.aspx
• The sessions
https://www.sqlsaturday.com/874/Sessions/SessionEvaluation.aspx

You might also like