Presentation is loading. Please wait.

Presentation is loading. Please wait.

Power BI Integration with Azure Machine Learning

Similar presentations


Presentation on theme: "Power BI Integration with Azure Machine Learning"— Presentation transcript:

1 Power BI Integration with Azure Machine Learning
Chad Dotzenrod Analytics & Cloud Evangelist at SWC Technology Partners

2 Meet Your Presenter: Chad Dotzenrod
Analytics & Cloud Evangelist at SWC Technology Partners PUG Membership: Greater Chicago Power BI User Group 20 Years of work in Data and Analytics Education: NDSU & DeVry Keller Fun facts! Avid fisherman and competitive BBQ contestant Follow me on Connect with me on LinkedIn: chaddotzenrod

3 Agenda Quick Audience Survey Why Is This Important?
Integration Scenarios Review Azure Machine Learning Workbench Demo Q&A

4 Audience Survey

5

6 Why is Azure Machine Learning Integration Important?
Start with the right Approach - Pop Culture Reference - Star Wars Death Star, etc

7 Making Decisions with Data
Descriptive > Predictive > Prescriptive Command and Control via Power BI Trust but Verify Understanding what the algorithms, models, and machines are doing to us! Use Power BI to question everything happening with the data Prove the ROI Show and Tell a story that is impactful to the organization Steer clear of Data Science jargon when presenting to citizens! Lift charts, AUC graphs, and False Positive rates make sense to data scientists and statisticians but not to the business

8 Microsoft TDSP Process Data Science is a process! No Shortcuts!
Team Data Science Process

9 Who is Solving What? Use Case Automotive Manufacturing Consumer
Finserve Energy Pharma Public Media Logistics Real-time Optimization Strategic Optimization Predictive Analytics Predictive Maintenance Radical Personalization Trend/Anomaly Discovery Forecasting Asking the Questions: You can’t deploy Power BI successfully if you only talk to IT and same goes if you only talk to the business Pragmatism: Know your users. Understand their journey when it comes to answering questions with data Competition: Are there other BI tools in place, Can the users break from Excel, Are there legacy reports that users depend on? Will they change Source: McKinsey – The Age of Analytics Report, Dec 2016

10 Integration Scenarios
House Marketing is fighting House Accounting, they don't see the White Walkers, you don't want the wall. WINTER IS COMING, WINTER IS HERE - One unified 7 Department Kingdoms…

11 Integration Scenarios
Plentiful Options Batch Real-Time On-Demand Complexity & Skillsets Latency/Data Prep/Cost SQL/R/Python/Azure

12

13

14 Review Azure Machine Learning Toolset
Security - Walking Dead –   Protect the wrong people from the wrong things… that's why we have role based security. Choose the right tool  - not a bow and arrow, you need a machete. Not just a firewall, you need xyz.

15 Azure Machine Learning Studio
Citizen Data Science Tool Visual Editor Gallery Support Reasonable Cost Pure PaaS

16 Azure Machine Learning Studio Details
What skillset is needed to get started? How much does it cost? Deployment Options Integration with Excel, LOB Apps, and Power BI Power BI is useful in telling the model story * Studio is meant for citizen data scientists and Workbench is for Professionals. Think of one as a visual IDE and the Other as a code based IDE. Both can get the same results but the latter has more options for configuration and customization First two users experiment accounts are free, You will be charged for model deployment compute if in cloud. Additional experiment accts are $50/month preview and likely $100/month GA

17 Azure Machine Learning Workbench
Professional Data Science Tool Enhanced Data Preparation Model Experimentation Model Management Bridges On-Premise, Hybrid, & Cloud

18 AML Workbench Details How is this different from AML PaaS? How much does it cost? Deployment Targets: Docker Containers (Local or Azure) Virtual Machines (Azure) SQL Server Machine Learning Server (Local or Azure)* Azure IoT Edge* Integration with Visual Studio AI Tools Excellent Companion Tool for Power BI Solution Builders! * Studio is meant for citizen data scientists and Workbench is for Professionals. Think of one as a visual IDE and the Other as a code based IDE. Both can get the same results but the latter has more options for configuration and customization First two users experiment accounts are free, You will be charged for model deployment compute if in cloud. Additional experiment accts are $50/month preview and likely $100/month GA *These deployment targets are not currently in preview yet but are on the roadmap

19 Power BI Challenges for AML Workbench
Data Profiling is not easy enough and many don’t know how to approach Working with “Big Data” can be cumbersome and slow Cleansing capabilities are excellent but not as rich as AMLW. No seamless Integration with GIT or other code repositories Power BI Adoption Speedbumps: Hard data challenges (complicated transforms) might have to go to the ETL Team Performance data challenges might go to Database Team (EDW, Cubes, In-memory) Functional challenges – I can’t do X with PBD and I don’t know how to write R Scripts * Studio is meant for citizen data scientists and Workbench is for Professionals. Think of one as a visual IDE and the Other as a code based IDE. Both can get the same results but the latter has more options for configuration and customization First two users experiment accounts are free, You will be charged for model deployment compute if in cloud. Additional experiment accts are $50/month preview and likely $100/month GA

20 Demo Demo/Overview of AML Workbench
Demo of Power BI integration with AML.

21 Thank You for Attending!
Don’t forget to join your local PUG to enjoy year-round networking and learning. Twitter: @cdotzenrod LinkedIn: chaddotzenrod


Download ppt "Power BI Integration with Azure Machine Learning"

Similar presentations


Ads by Google