Deployment and Governance of Power BI for the

Deployment and Governance of Power BI for the Enterprise Zach Willand

About Me • Business Intelligence Developer • Davis H. Elliot • Microsoft shop • Power Platform • 46 Solutions • Data Driven • Research Analyst • Assets Protection • IT • SQL, Power. BI, Azure, Power. Shell

Overview • Data Architecture • Data Modeling • Architecture Considerations • Import vs Direct Query vs Live • Data Flows / Data Sets • Design • Leveraging Bookmarks and Button Navigation • Mobile • Deployment Model • Workspace vs Apps • Apps Deployment • Permissions Deployment Design Data Architecture

Data Model / Architecture

Data Model / Architecture • Real Time vs. Near Real Time • Costly • Often 90% more work for 10% more benefit • May be impossible in many cases • Near Real Time • Much more cost effective • In Power. BI, near real time is typically around 2 hours • Perfect for MOST scenarios

Data Model / Architecture • Import (Cached) • Local, in memory, fast • Near-Real Time – Scheduled Refresh • Much more flexible • All data sources available • Full Power. BI Capabilities • Mash up between data sources • 1 gb pro 10 gb prem

Data Model / Architecture • Direct Query • • • Real Time Limitations Data stays in data source SLOW – performance issues Limited amount of data sources • Relational Databases • Limited Power Query / DAX • Time intelligence = gone

Data Model / Architecture • Live connection • • SSAS Only Modeling must be done in SSAS No data mashup in PBI Supports much larger model size

Data Model / Architecture • Use Import mode when possible • Use Query Editor when possible for performance (report-wide filtering for example) • Avoid sources like Share. Point for large datasets • I. E. Offloading data mapping to others • If data has not yet been collected, plan for scalability • Power. BI Service WILL timeout from slower sources with large data sets Data Architecture

Data Flows

Data Flows Source Data. Flows Datasets Report/App

Data Flows Benefits • Allows for different refresh times for data sources in single report • Power. Query / ETL in cloud • Allows for snapshots of data • I. E. One dataset per quarter • Date table never needs refresh, speeding refresh time • Only hit the data source ONE TIME, even with multiple reports refreshing • Much faster dataset refresh times

Data Flows

Data Flows • Multiple data sources, types in each Data. Flow container

Data Flows Summary / Refresh Method Source Data. Flows Datasets Report/App Dataset

Data Flows • Can you migrate an already established report to a Data. Flow? • YES

Data Sets • • Much more limited in scope Larger orgs Locked in Based on already created report datasets • End user can access if they have workspace permissions • One reason to use Apps instead of Workspaces

Data Sets • No Data Model • Cannot add data sources • No calc. columns • No Query editor

Data Model / Architecture • Keys to Success: • Data modelling is one of the most important skills • Tutorials are ok in the beginning • Move to real data, real problems • Power BI will allow ‘bad’ data models • DAX • Data model first: bad model, bad numbers

Report Design • Leverage bookmarks and buttons to enhance user experience, usable canvas area and to take your report to the next level • Create mobile views for KPIs and other key metrics Design Data Architecture

Report Design

Deployment • Workspaces vs. Apps • Keeping App users up to date Deployment Design Data Architecture

Deployment • Workspaces: • • Typically thought of as developer environments Changes appear immediately. Useful for collaboration, testing Allows users to re-use datasets • Apps: • End-user environments • Isolated • No access to backend datasets (but users can still export data if allowed) • Changes do not deploy until app is updated manually • Can have multiple reports in a single app

Deployment App Permissions Workspace Permissions

Deployment

Deployment • Keeping App users up to date: • Updates page • Fix Up page

Deployment • Leverage Power. Shell • Power BI API / Module • Refreshing • Monitoring • Azure Automation

Questions? @zachwilland zwilland@dhec. com