Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
SlideShare a Scribd company logo

Afternoons with Azure - Power BI and Azure Analysis Services

•Download as PPTX, PDF•
•
CCG
CCG

See how Microsoft Power BI and Azure Analysis Services are influencing the BI and analytics market. Journey through data structures and fundamentals for setting up your next dashboard initiative. Interested in learning more? Click ccganalytics.com/resources for more or call (813) 265-3239.

1 of 66
Afternoons with Azure Series Microsoft Power BI and Azure Analysis Services
Intro to CCG Azure Subscriptions Overview of Power BI Overview of Azure Analysis Services Power BI Workshop Q & A
Microsoft Power BI Martin Rivera, BI Architect
Overview of Power BI Power BI Desktop Data Sources Visualizations Interactive Reports
Overview of Power BI
Microsoft in the BI and Analytics Market Leaders Leaders are vendors that demonstrate a solid understanding of the product capabilities and commitment to customer success that buyers demand in the current market. Have a robust roadmap for solidifying its position as a future market leader, thus protecting the investment of today's buyers. Challengers Challengers are well-positioned to succeed in the market. However, they may be limited to specific use cases, technical environments or application domains. Visionaries Visionaries have a strong and unique vision for delivering a modern analytics and BI platform. However, they may have gaps relating to broader functionality, customer experiences, operations, and sales execution. Niche Players Niche Players do well in a specific segment of the analytics and BI market — such as cloud BI, customer-facing analytics, agile reporting and dashboarding, embeddability or big data analytics — or have a limited capability to out-innovate or outperform other vendors. Microsoft has been leader for 11 years
Mi What is Power BI? Power BI is a suite of business analytics tools to analyze data and share insights. Monitor your business and get answers quickly with rich dashboards available on every device. Microsoft BI Stack
What is Power BI Report Server Power BI Report Server (PBIRS), is a Microsoft solution deployed on premise for creating, publishing, and managing reports, then delivering them to the right users in different ways. For end users of Power BI Report Server, the front door is a modern web portal you can view in any web browser. *Content gets developed and published using the Power BI Desktop Tool. Invoices Time Vendor State Customer
What’s the Difference? Power BI (Service) • Hosted in the (Azure) cloud • SaaS (Software as a Service) • Integration with Azure stack • Rapid evolution (~monthly) • Cutting edge features (ex: Cortana) Power BI Report Server • Hosted on premise • Rely on legacy SSRS architecture • Integration with Microsoft tools • More controlled rollouts • Features backlogged at times Both aim to provide developers and end-users with a similar experience.
Report Power BI Dashboard A Power BI report displays one or more visualizations that represent different findings and insights from a specific dataset. Could be a single page, or multiple pages. All visualizations displayed are part of the same report. A Power BI dashboard is a single page, often called a canvas, that uses visualizations to tell a story. The visualizations you see on the dashboard are called tiles and are pinned to the dashboard from existing reports. Typically, a PBI dashboard will contain tiles from various reports. What’s the Difference?
Power BI Desktop
Power BI Desktop • Intuitive report Authoring • Data Preparation and Data Modeling • Drag-and-drop to place content exactly where you want it • Discovery of data patterns in one unified view through rich and interactive visualizations NOTE: There are separate versions of Power BI Desktop for Power BI Service and Power BI Report Service
Power BI Desktop Publish Report Server
Power BI Desktop Clean and Transform Your Data Query Editor • 1 query per table/worksheet • Transform existing columns • Add new columns • Combine multiple queries • Merge • Append
Power BI Desktop Clean and Transform Your Data Data Type Specific Filters Replace/Fill Values
Power BI Desktop Modeling • The Semantic Model • Friendly names • Business rules / definitions • Creating custom columns & measures • Formatting columns and measures for reporting • E.g. currency, number of decimals, short/long date
Power BI Desktop • Can be configured Between any two tables, even from different sources • Columns must have the same data type • Column names don’t have to be the same • Supports 1 to 1, 1 to Many, and Many to 1 relationships • Many-to many relationship are not supported: use a bridge table Relationships between Entities
Power Query “M” Language • Similar to MDX • Often used with multidimensional models • Mechanism for manipulating data prior to load • Can be a powerful tool and extend capabilities available from UI https://msdn.microsoft.com/en-us/library/mt211003.aspx Power Query “M” Language Reference
Data Analysis Expressions (DAX) • Calculated Columns – Calculated at load time – Evaluated for each row – Stored in the model! • Calculated Measures – Calculated at (report) run time – Evaluated with context (filters) – Implicit: Created by Power BI – Explicit: Created by you • Calculated Tables – Lookups – Date Table – Constants • DAX is a library of functions and operators that can be used after you’ve loaded your dataset • Conceptually similar to Excel formulas, but different in how it may process inputs • Concepts are straightforward, but not easy – Don’t expect to learn DAX in 2 hours • Functional Language – Full code in nested functions – Evaluated from inner to outer • Do not write everything on one line Over 200 Functions: • Text Functions • Statistical Functions • Math and Trig Functions • Date and Time • Information Functions • Filter Functions
Data Sources
Types of Data Sources
Data Connectivity Mode This is handled in Data Source Settings in Power BI Desktop. The report consumer or content manager does not have the option to control/override this. Import Imports data into the generated PBIX file. This results in faster performance, but is a “snapshot” of data as of a point in time*. *You can schedule a periodic refresh. DirectQuery Connects to the data source “live” at each report execution. This means data is in real-time, but can result in slower performance of report executions and interactions. You cannot mix both data connectivity modes in the same report You cannot change the data connectivity modes of a report once selected
Visualizations
• Power BI is all about visualizations • Tell the right story • Highlight unforeseen nuances • Easy to consume • Intuitive, and actionable Visualize Data for Maximum Impact
Default Visualizations 1. Stacked Bar Chart 2. Stacked Column chart 3. Clustered Bar Chart 4. Clustered Column Chart 5. 100% Stacked Bar Chart 6. 100% Stacked Column chart 7. Line Chart 8. Area Chart 9. Stacked Area Chart 10. Line and Stacked Column Chart 11. Line and Clustered Column Chart 12. Ribbon Chart 13. Waterfall Chart 14. Scatter Chart 15. Pie Chart 16. Donut Chart 17. Tree Map 18. Map 19. Filled Map 20. Funnel 21. Gauge 22. Card 23. Multi-Row Card 24. KPI 25. Slicer 26. Table 27. Matrix 28. R Script Visual
Custom Visuals
Custom Visuals • Available from Microsoft Apps site • They are Open Source • Users contribute, MSFT publishes • They are someone else’s idea of what looks good • May provide limited properties/controls • Can also be obtained directly from other publishers • No guarantee they will work! • May not be supported across versions • Reside within the specific PBIX file Example: Body Part Analysis
Interactive Reports
Interactive Analysis Year Selected State Selected All visualizations interact with one another. Some visualizations are filtered, while others are highlighted Only visualizations within the same page interact with one another. Interactions can be controlled within Power BI Desktop
Drilling Power BI allows for drilling down and up hierarchies to explore data at different levels of a pre-defined hierarchy • Great use case for exploring time based hierarchies Date columns have built-in Time Intelligence, to generate hierarches Also, you have the ability to setup drill through filters on a report page, to focus the data in that child page, based on the selection in the parent page Drill through can be setup between pages w/in the same report Year Quarter Month Day
Drill Down vs. Drill Through
Intro to Azure Analysis Services
SSAS Fundamentals – Multidimensional Models – Tabular Models Azure Analysis Services – Features – Compatibilities with On Premises SSAS – Deployment Why Azure?
SSAS Fundamentals
The Model, regardless of type, will be the Semantic Layer for the analytics user SSAS Model Types Multidimensional Model Tabular Model
Online Analytical Processing (Traditional OLAP) Cubes consist of dimensions and measures Create cubes with SSDT, manage with SSDT and SSMS Default Mode (MOLAP) stores imported data on disk (Requires large volume of disk space) Performance is increased by pre-calculating and pre-loading aggregations Query data via SSMS, MS Excel (most common), SSRS, Power BI and other BI Tools SSAS MDM is most commonly queried using MDX (Multidimensional eXpressions) SSAS server must be configured in Multidimensional Mode, a separate instance will be needed for Tabular Models SSAS Multidimensional Model
Multidimensional Model Features Aggregations Calculated Measures Drill Through Drill Down Hierarchies KPI’s Many-to-many relationships Partitions Perspectives Row and Object Level Security
Multidimensional Data Sources Any on premises data source – Oracle – IBM – SAP – Teradata – Microsoft (Access, SQL Server, Excel, etc.) – Any data source with OLEDB (OLAP sources not supported)
Multidimensional Cube Structure Yellow – Fact Table Blue – Dimension Table
Sample Dimension Usage
MDM Sample Calculation (used in aggregations)
Partitions and Perspectives • Perspectives (Logical views for organization for end users) For example: Subject Area • Partitions (Slices or chunks of data optimized for loading data)
While query performance of cubes is very fast for large datasets, Loading time can be very long. The cube is unavailable while loading. Dimensions and aggregations will need to be recalculated every time the cube is processed. • Partitions allow for horizontal slices of data to be loaded (By Year, Geography, Customer, etc.) • Incremental partitions can be used to load the most recent data (more often), with historical partitions loaded less often. Partitions
Perspectives provide horizontal slices (views) of the data in the cube Cubes CANNOT be secured at the Perspective level Perspectives
Performance enhanced by predefined aggregations The development to production cycle is longer than most solutions Multidimensional data structures are rigid. Must be loaded as facts and dimensions. Flattened summary views not supported for loading Default storage is MOLAP. Direct Query is available (ROLAP), but does not utilize pre-calculated and loaded aggregations Many companies use a hybrid approach (HOLAP). Larger facts using MOLAP, smaller facts using ROLAP. Not included in Azure Analysis Services SSAS Multidimensional Model Takeaways
Overview SSAS Tabular Model A Tabular Model is a relational, columnar, multi- threaded SSAS database that utilizes the xVelocity (Vertipaq) compression technology in the following modes: • In Memory (Default) • Direct Query
Design Design, Manage, Explore Manage Explore
Metadata The Tabular Model utilizes the relationships inherited from its data sources, where applicable.
Data Compression Multidimensional models – 3x compression Tabular models – 10x compression
Feature In Memory Direct Query MDM Calculated Column Yes No No Calculated Measure Yes Yes Yes Calculated Table Yes No No Hierarchies Yes Yes Yes, more flexibility KPI’s Yes Yes Yes Partitions Yes Yes Yes Perspectives Yes Yes Yes MDM type Aggregations No No Yes Multiple data sources Yes No Yes Row Level Security Yes (Row Filter) Only from database Yes Data Modification Language (DML) DAX, MDX DAX MDX, DMX SSAS Tabular Model Features
https://docs.microsoft.com/en-us/sql/analysis-services/comparing-tabular-and-multidimensional-solutions-ssas Feature Comparison (Complete List)
Data Source In Memory Direct Query Multiple Yes No Relational DB Yes Yes MDM Cube Yes No Spreadsheet Yes No Text Yes No Data Feeds Yes No Azure SQL Database Yes Yes Many-to-many relationships Requires Bi-directional Cross filter and DAX Requires Bi-directional Cross filter and DAX Tabular Model Data Sources
Best performance/compression Partitions load in parallel Data limited by installed memory, also no single column in a table can have more than 2 billion distinct values. Full DAX capabilities DAX Time Intelligence DAX or MDX Multiple data sources Tabular In Memory vs Direct Query Direct Query Only stores metadata (with exception of sample partitions) Does not consume memory for storage Useful for instances without large memory, no data size limitation Does not support DAX calculated columns or tables Calculated measures are supported DAX, but not Time Intelligence Single data source (Process other data sources through SQL Server) Inherits Windows AD RLS (may require Kerberos) In Memory
SSAS Tabular Model Relational View
SSAS Tabular Model Data View with Measures
SSAS Tabular KPI’s Simple interface 1. Base Measure 2. Target (Measure or literal value) 3. Target ranges 4. Icon style
Azure Analysis Services
Coverage Specs Model(s) Tabular Model >= 1200 Compatibility Design Tool SSDT, Web Designer (Preview) SSMS to manage Deployment SSDT Data Sources Same as SSAS Tabular and more Local sources require on premises data gateway Features Fully compatible with SSAS Tabular Modes In Memory, Direct Query Azure Analysis Services – Overview
Azure data source In-memory Direct Query Azure SQL Database Yes Yes Azure SQL Data Warehouse Yes Yes Azure Blob Storage Yes No Azure Table Storage Yes No Azure Cosmos DB Yes No Azure Data Lake Store Yes No Azure HDInsight HDFS Yes No Azure HDInsight Spark (Beta) Yes No Azure Analysis Services - Cloud Data Sources (1400 Compatibility)
1. Login to Azure Portal to copy destination server name 2. In SSDT (Solution Explorer), right-click the Project > Properties 3. Under Deployment paste server name 4. In Solution Explorer, right-click Properties, then select Deploy (You may be prompted to sign into Azure) Steps to Deploy On-prem SSAS Tabular to Azure Analysis Services
Task Tool Manage SSMS Design SSDT, Web Designer Explore Excel, Power BI Desktop, SSRS?, Any supported BI Tool Interacting with Azure Analysis Services
Why Azure?
Better with Azure Azure Integration Azure Active Directory Azure Data Factory Azure Automation and Functions Rapid Development/Deployment Create server in minutes Model data in SSDT or Web Designer Import model from PBIX file Scalable Tier based support (price based on needs) Use query pools to replicate models to improve client performance Use geographical Azure regions to minimize latency. Deploy in multiple regions for high availability Built on SSAS Foundation Familiar tools (SSMS, SSDT) Compatible with SSAS Tabular (>=1200) TMSL compatible (Scripting Language) Existing data sources + cloud
Power BI Workshop
Afternoons with Azure - Power BI and Azure Analysis Services
THANK YOU!

More Related Content

Afternoons with Azure - Power BI and Azure Analysis Services

  • 1. Afternoons with Azure Series Microsoft Power BI and Azure Analysis Services
  • 2. Intro to CCG Azure Subscriptions Overview of Power BI Overview of Azure Analysis Services Power BI Workshop Q & A
  • 3. Microsoft Power BI Martin Rivera, BI Architect
  • 4. Overview of Power BI Power BI Desktop Data Sources Visualizations Interactive Reports
  • 6. Microsoft in the BI and Analytics Market Leaders Leaders are vendors that demonstrate a solid understanding of the product capabilities and commitment to customer success that buyers demand in the current market. Have a robust roadmap for solidifying its position as a future market leader, thus protecting the investment of today's buyers. Challengers Challengers are well-positioned to succeed in the market. However, they may be limited to specific use cases, technical environments or application domains. Visionaries Visionaries have a strong and unique vision for delivering a modern analytics and BI platform. However, they may have gaps relating to broader functionality, customer experiences, operations, and sales execution. Niche Players Niche Players do well in a specific segment of the analytics and BI market — such as cloud BI, customer-facing analytics, agile reporting and dashboarding, embeddability or big data analytics — or have a limited capability to out-innovate or outperform other vendors. Microsoft has been leader for 11 years
  • 7. Mi What is Power BI? Power BI is a suite of business analytics tools to analyze data and share insights. Monitor your business and get answers quickly with rich dashboards available on every device. Microsoft BI Stack
  • 8. What is Power BI Report Server Power BI Report Server (PBIRS), is a Microsoft solution deployed on premise for creating, publishing, and managing reports, then delivering them to the right users in different ways. For end users of Power BI Report Server, the front door is a modern web portal you can view in any web browser. *Content gets developed and published using the Power BI Desktop Tool. Invoices Time Vendor State Customer
  • 9. What’s the Difference? Power BI (Service) • Hosted in the (Azure) cloud • SaaS (Software as a Service) • Integration with Azure stack • Rapid evolution (~monthly) • Cutting edge features (ex: Cortana) Power BI Report Server • Hosted on premise • Rely on legacy SSRS architecture • Integration with Microsoft tools • More controlled rollouts • Features backlogged at times Both aim to provide developers and end-users with a similar experience.
  • 10. Report Power BI Dashboard A Power BI report displays one or more visualizations that represent different findings and insights from a specific dataset. Could be a single page, or multiple pages. All visualizations displayed are part of the same report. A Power BI dashboard is a single page, often called a canvas, that uses visualizations to tell a story. The visualizations you see on the dashboard are called tiles and are pinned to the dashboard from existing reports. Typically, a PBI dashboard will contain tiles from various reports. What’s the Difference?
  • 12. Power BI Desktop • Intuitive report Authoring • Data Preparation and Data Modeling • Drag-and-drop to place content exactly where you want it • Discovery of data patterns in one unified view through rich and interactive visualizations NOTE: There are separate versions of Power BI Desktop for Power BI Service and Power BI Report Service
  • 14. Power BI Desktop Clean and Transform Your Data Query Editor • 1 query per table/worksheet • Transform existing columns • Add new columns • Combine multiple queries • Merge • Append
  • 15. Power BI Desktop Clean and Transform Your Data Data Type Specific Filters Replace/Fill Values
  • 16. Power BI Desktop Modeling • The Semantic Model • Friendly names • Business rules / definitions • Creating custom columns & measures • Formatting columns and measures for reporting • E.g. currency, number of decimals, short/long date
  • 17. Power BI Desktop • Can be configured Between any two tables, even from different sources • Columns must have the same data type • Column names don’t have to be the same • Supports 1 to 1, 1 to Many, and Many to 1 relationships • Many-to many relationship are not supported: use a bridge table Relationships between Entities
  • 18. Power Query “M” Language • Similar to MDX • Often used with multidimensional models • Mechanism for manipulating data prior to load • Can be a powerful tool and extend capabilities available from UI https://msdn.microsoft.com/en-us/library/mt211003.aspx Power Query “M” Language Reference
  • 19. Data Analysis Expressions (DAX) • Calculated Columns – Calculated at load time – Evaluated for each row – Stored in the model! • Calculated Measures – Calculated at (report) run time – Evaluated with context (filters) – Implicit: Created by Power BI – Explicit: Created by you • Calculated Tables – Lookups – Date Table – Constants • DAX is a library of functions and operators that can be used after you’ve loaded your dataset • Conceptually similar to Excel formulas, but different in how it may process inputs • Concepts are straightforward, but not easy – Don’t expect to learn DAX in 2 hours • Functional Language – Full code in nested functions – Evaluated from inner to outer • Do not write everything on one line Over 200 Functions: • Text Functions • Statistical Functions • Math and Trig Functions • Date and Time • Information Functions • Filter Functions
  • 21. Types of Data Sources
  • 22. Data Connectivity Mode This is handled in Data Source Settings in Power BI Desktop. The report consumer or content manager does not have the option to control/override this. Import Imports data into the generated PBIX file. This results in faster performance, but is a “snapshot” of data as of a point in time*. *You can schedule a periodic refresh. DirectQuery Connects to the data source “live” at each report execution. This means data is in real-time, but can result in slower performance of report executions and interactions. You cannot mix both data connectivity modes in the same report You cannot change the data connectivity modes of a report once selected
  • 24. • Power BI is all about visualizations • Tell the right story • Highlight unforeseen nuances • Easy to consume • Intuitive, and actionable Visualize Data for Maximum Impact
  • 25. Default Visualizations 1. Stacked Bar Chart 2. Stacked Column chart 3. Clustered Bar Chart 4. Clustered Column Chart 5. 100% Stacked Bar Chart 6. 100% Stacked Column chart 7. Line Chart 8. Area Chart 9. Stacked Area Chart 10. Line and Stacked Column Chart 11. Line and Clustered Column Chart 12. Ribbon Chart 13. Waterfall Chart 14. Scatter Chart 15. Pie Chart 16. Donut Chart 17. Tree Map 18. Map 19. Filled Map 20. Funnel 21. Gauge 22. Card 23. Multi-Row Card 24. KPI 25. Slicer 26. Table 27. Matrix 28. R Script Visual
  • 27. Custom Visuals • Available from Microsoft Apps site • They are Open Source • Users contribute, MSFT publishes • They are someone else’s idea of what looks good • May provide limited properties/controls • Can also be obtained directly from other publishers • No guarantee they will work! • May not be supported across versions • Reside within the specific PBIX file Example: Body Part Analysis
  • 29. Interactive Analysis Year Selected State Selected All visualizations interact with one another. Some visualizations are filtered, while others are highlighted Only visualizations within the same page interact with one another. Interactions can be controlled within Power BI Desktop
  • 30. Drilling Power BI allows for drilling down and up hierarchies to explore data at different levels of a pre-defined hierarchy • Great use case for exploring time based hierarchies Date columns have built-in Time Intelligence, to generate hierarches Also, you have the ability to setup drill through filters on a report page, to focus the data in that child page, based on the selection in the parent page Drill through can be setup between pages w/in the same report Year Quarter Month Day
  • 31. Drill Down vs. Drill Through
  • 32. Intro to Azure Analysis Services
  • 33. SSAS Fundamentals – Multidimensional Models – Tabular Models Azure Analysis Services – Features – Compatibilities with On Premises SSAS – Deployment Why Azure?
  • 35. The Model, regardless of type, will be the Semantic Layer for the analytics user SSAS Model Types Multidimensional Model Tabular Model
  • 36. Online Analytical Processing (Traditional OLAP) Cubes consist of dimensions and measures Create cubes with SSDT, manage with SSDT and SSMS Default Mode (MOLAP) stores imported data on disk (Requires large volume of disk space) Performance is increased by pre-calculating and pre-loading aggregations Query data via SSMS, MS Excel (most common), SSRS, Power BI and other BI Tools SSAS MDM is most commonly queried using MDX (Multidimensional eXpressions) SSAS server must be configured in Multidimensional Mode, a separate instance will be needed for Tabular Models SSAS Multidimensional Model
  • 37. Multidimensional Model Features Aggregations Calculated Measures Drill Through Drill Down Hierarchies KPI’s Many-to-many relationships Partitions Perspectives Row and Object Level Security
  • 38. Multidimensional Data Sources Any on premises data source – Oracle – IBM – SAP – Teradata – Microsoft (Access, SQL Server, Excel, etc.) – Any data source with OLEDB (OLAP sources not supported)
  • 39. Multidimensional Cube Structure Yellow – Fact Table Blue – Dimension Table
  • 41. MDM Sample Calculation (used in aggregations)
  • 42. Partitions and Perspectives • Perspectives (Logical views for organization for end users) For example: Subject Area • Partitions (Slices or chunks of data optimized for loading data)
  • 43. While query performance of cubes is very fast for large datasets, Loading time can be very long. The cube is unavailable while loading. Dimensions and aggregations will need to be recalculated every time the cube is processed. • Partitions allow for horizontal slices of data to be loaded (By Year, Geography, Customer, etc.) • Incremental partitions can be used to load the most recent data (more often), with historical partitions loaded less often. Partitions
  • 44. Perspectives provide horizontal slices (views) of the data in the cube Cubes CANNOT be secured at the Perspective level Perspectives
  • 45. Performance enhanced by predefined aggregations The development to production cycle is longer than most solutions Multidimensional data structures are rigid. Must be loaded as facts and dimensions. Flattened summary views not supported for loading Default storage is MOLAP. Direct Query is available (ROLAP), but does not utilize pre-calculated and loaded aggregations Many companies use a hybrid approach (HOLAP). Larger facts using MOLAP, smaller facts using ROLAP. Not included in Azure Analysis Services SSAS Multidimensional Model Takeaways
  • 46. Overview SSAS Tabular Model A Tabular Model is a relational, columnar, multi- threaded SSAS database that utilizes the xVelocity (Vertipaq) compression technology in the following modes: • In Memory (Default) • Direct Query
  • 48. Metadata The Tabular Model utilizes the relationships inherited from its data sources, where applicable.
  • 49. Data Compression Multidimensional models – 3x compression Tabular models – 10x compression
  • 50. Feature In Memory Direct Query MDM Calculated Column Yes No No Calculated Measure Yes Yes Yes Calculated Table Yes No No Hierarchies Yes Yes Yes, more flexibility KPI’s Yes Yes Yes Partitions Yes Yes Yes Perspectives Yes Yes Yes MDM type Aggregations No No Yes Multiple data sources Yes No Yes Row Level Security Yes (Row Filter) Only from database Yes Data Modification Language (DML) DAX, MDX DAX MDX, DMX SSAS Tabular Model Features
  • 52. Data Source In Memory Direct Query Multiple Yes No Relational DB Yes Yes MDM Cube Yes No Spreadsheet Yes No Text Yes No Data Feeds Yes No Azure SQL Database Yes Yes Many-to-many relationships Requires Bi-directional Cross filter and DAX Requires Bi-directional Cross filter and DAX Tabular Model Data Sources
  • 53. Best performance/compression Partitions load in parallel Data limited by installed memory, also no single column in a table can have more than 2 billion distinct values. Full DAX capabilities DAX Time Intelligence DAX or MDX Multiple data sources Tabular In Memory vs Direct Query Direct Query Only stores metadata (with exception of sample partitions) Does not consume memory for storage Useful for instances without large memory, no data size limitation Does not support DAX calculated columns or tables Calculated measures are supported DAX, but not Time Intelligence Single data source (Process other data sources through SQL Server) Inherits Windows AD RLS (may require Kerberos) In Memory
  • 54. SSAS Tabular Model Relational View
  • 55. SSAS Tabular Model Data View with Measures
  • 56. SSAS Tabular KPI’s Simple interface 1. Base Measure 2. Target (Measure or literal value) 3. Target ranges 4. Icon style
  • 58. Coverage Specs Model(s) Tabular Model >= 1200 Compatibility Design Tool SSDT, Web Designer (Preview) SSMS to manage Deployment SSDT Data Sources Same as SSAS Tabular and more Local sources require on premises data gateway Features Fully compatible with SSAS Tabular Modes In Memory, Direct Query Azure Analysis Services – Overview
  • 59. Azure data source In-memory Direct Query Azure SQL Database Yes Yes Azure SQL Data Warehouse Yes Yes Azure Blob Storage Yes No Azure Table Storage Yes No Azure Cosmos DB Yes No Azure Data Lake Store Yes No Azure HDInsight HDFS Yes No Azure HDInsight Spark (Beta) Yes No Azure Analysis Services - Cloud Data Sources (1400 Compatibility)
  • 60. 1. Login to Azure Portal to copy destination server name 2. In SSDT (Solution Explorer), right-click the Project > Properties 3. Under Deployment paste server name 4. In Solution Explorer, right-click Properties, then select Deploy (You may be prompted to sign into Azure) Steps to Deploy On-prem SSAS Tabular to Azure Analysis Services
  • 61. Task Tool Manage SSMS Design SSDT, Web Designer Explore Excel, Power BI Desktop, SSRS?, Any supported BI Tool Interacting with Azure Analysis Services
  • 63. Better with Azure Azure Integration Azure Active Directory Azure Data Factory Azure Automation and Functions Rapid Development/Deployment Create server in minutes Model data in SSDT or Web Designer Import model from PBIX file Scalable Tier based support (price based on needs) Use query pools to replicate models to improve client performance Use geographical Azure regions to minimize latency. Deploy in multiple regions for high availability Built on SSAS Foundation Familiar tools (SSMS, SSDT) Compatible with SSAS Tabular (>=1200) TMSL compatible (Scripting Language) Existing data sources + cloud