Presentation is loading. Please wait.

Presentation is loading. Please wait.

Business Intelligence Fundamentals: Unified Dimensional Model

Similar presentations


Presentation on theme: "Business Intelligence Fundamentals: Unified Dimensional Model"— Presentation transcript:

1 Business Intelligence Fundamentals: Unified Dimensional Model
9/17/2018 Business Intelligence Fundamentals: Unified Dimensional Model Ola Ekdahl IT Mentors

2 Agenda Reviewing OLAP Fundamentals Understanding Database Components
Business Intelligence Fundamentals: Unified Dimensional Model Reviewing OLAP Fundamentals Understanding Database Components Programming SSAS Agenda

3 Where Are We? Data Warehouse Data Sources Data Marts Staging Area
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Where Are We? Data Warehouse Data Sources Data Marts Staging Area Manual Cleansing

4 Reviewing OLAP Fundamentals
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Reviewing OLAP Fundamentals Introduction to OLAP Business Scenarios SQL Server 2008 Analysis Services Unified Dimensional Model (UDM) Benefits of the UDM

5 Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Introduction to OLAP Contains structures optimized for rapid ad hoc information retrieval Includes a calculation engine for fast, flexible transformation of base data Supports discovery of business trends and statistics not directly visible in data warehouse queries OLAP = OnLine Analytical Processing This slide reviews the OLAP concepts covered in Module 01. Contrast OLAP with OLTP systems which are optimized for data entry, not for reporting. Also contrast with relational data warehouse design which optimizes for reporting but does not easily support trending and statistics and requires maintenance of summary tables for high performance.

6 Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model What sales did we expect to achieve in North America for CY 2004 Q1? Querying the Cube 5,005,000 Use this animated slide to demystify how results are retrieved from cubes. Point out that in reality cubes are more than just three dimensions; this example is greatly simplified.

7 Business Scenarios Sales analysis
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Business Scenarios Sales analysis Budgeting and forecast collection and reporting Financial reporting Web statistics analysis Survey results analysis ETL process analysis Can apply to any scenario involving aggregated data that needs to be queried fast and flexibly at different levels of summarization Use this slide to highlight common business scenarios for OLAP. Take care to point out that these scenarios are not just private-sector focused – they are equally relevant to government organizations. Examples: Sales analysis: Answer questions such as “How many widgets did we sell?” “What’s the revenue of a particular product line?” and “How is the sales staff performing?” Budgeting and forecast collection and reporting: Perform calculations by departments and/or cost and profit centers within an organization. This is the time to briefly discuss the potential for cube writeback (it is not discussed otherwise in this module). Financial reporting: Do general ledger reporting by ledger and time period, with a rich layer of calculations to support financial reporting based on YTD, ROI, and so on. Web statistics analysis: Analyze an Internet server log file, which is multidimensional by nature. Its facts can be various counters: visitors count, hits count, and other information contained in a log. Survey results analysis: Analyze the results of marketing surveys and investigations. ETL process analysis: ETL logging can represent vast volumes of data easily understood through multidimensional analysis. Stress the highlighted point on this slide: OLAP scenarios are relevant for any requirement to query data fast and flexibly–particularly for aggregation results. OLAP lends itself well to queries that often use the word “by”. For example, “I need to know sales BY calendar quarter BY region BY product line BY etc….” The repetition of this tiny word in a request is frequently a hint that the information could be efficiently retrieved from a multidimensional structure.

8 SQL Server 2008 Analysis Services
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model SQL Server 2008 Analysis Services OLAP component Aggregates and organizes data from business data sources Performs calculations difficult to perform using relational queries Supports advanced business intelligence, such as Key Performance Indicators Data mining component Discovers patterns in both relational and OLAP data Enhances the OLAP component with discovered results Use this slide to focus on SSAS’s OLAP component. Remind students that data mining is the second component of SSAS. Module 08 introduces data mining concepts.

9 Unified Dimensional Model (UDM)
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Unified Dimensional Model (UDM) Provides a bridge between the end user and the data sources Consolidates business rules into a single model Challenges traditional models Commonly called a Cube in SSAS 2000 Mention that SSAS 2005 introduced the UDM to blur the line between relational and multidimensional OLAP databases and to combine the best aspects of traditional OLAP analysis and relational reporting. The cubes and dimensions defined in SSAS are presented as a single, unified dimension model because they cover and serve the needs of both multidimensional OLAP analysis and relational reporting. The UDM facilitates isolation from heterogeneous back-end data sources and delivers improved performance for summary-type queries. Keep the discussion here very high-level and focused on what the UDM has been designed to achieve; the benefits of the UDM are listed on the following slide. Consider using examples that are common in many organizations today where data is accessed directly or via a variety of different models. Clearly, that approach works against the “one version of the truth” philosophy that the UDM upholds. Discussion could lead to how practical it may be to have a single UDM in the enterprise. The conclusion for some organizations will be that it is a nice theory but not terribly practical. Although the UDM supports relational-style reporting, you will often obtain best performance by querying the relational data marts directly.

10 Benefits of the UDM Permits extensive enhancements to the user model
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Benefits of the UDM Permits extensive enhancements to the user model Provides high-performance queries supporting interactive analysis, even over large volumes of data Supports richer analysis by encapsulating business rules Can abstract multiple data sources Supports “closing the loop,” letting users act on data The user model can be enriched with hierarchies, calculations, KPIs, actions, perspectives, and translations. These will be introduced and discussed later in this module. The second bullet point encapsulates the very definition of OLAP. Business rules (calculations) can be defined once and used consistently across all queries and reports. “Closing the loop” is supported by actions that will be discussed later.

11 Building an Analysis Services Database
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Building an Analysis Services Database DEMO Refer to the demonstration notes.

12 Understanding Database Components
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Understanding Database Components Data Source Data Source View Cube Dimensions Calculations Key Performance Indicators This slide presents the general recipe for creating an SSAS database. Do not spend time discussing these in detail; this slide just introduces the following slides.

13 Data Source Stores a connection string for a data source
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Data Source Stores a connection string for a data source OLE DB Provider Managed .NET Provider Supports SQL Server and other popular databases The Data Source object is simply a definition of where to find the data. Note that there is no requirement that SSAS consume any data from the SQL Server database engine.

14 Data Source View Metadata definition of schema elements
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Data Source View Metadata definition of schema elements Invisible source for client applications Benefits Focuses on a subset of tables in the data source(s) Allows integration of multiple sources Requires only read access to the underlying definitions and data Enables designing when disconnected from the data source(s) Insulates objects from changes in the underlying data source: The concept of data source views will be new to SSAS 2000 developers. The concept and the benefits should have been conveyed during Demo 1. For the last point, explain that dimensions, cubes, data mining models and report models MUST be based on data source views. So long as the interface between the DSV and the object does not change, then the developer has flexibility in managing the interface between the data source and the DSV. Explain that the data source view also has relevance to SSIS and Report Model projects. However, the second benefit (Allows integration of multiple sources) does not apply in these projects. Also, data source views in SSIS are optional.

15 Cube Combination of measures and dimensions as one conceptual model
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Cube Combination of measures and dimensions as one conceptual model Measures are sourced from fact tables Dimensions are sourced from dimension tables Rich data model enhanced by Calculations Key Performance Indicators (KPIs) Actions Perspectives Translations Partitions For beginners, it is often useful to explain that while a cube has three dimensions, cubes in SSAS can have any number of dimensions. The term cube is synonymous with “multidimensional” database. Discuss Actions, Perspectives, and Translations in detail at this point because they are not covered again. The remaining features will be addressed by specific slides later in the module. Mar-2008

16 Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Dimensions Create members from one or more tables or views in a data source view Contain attributes that correspond to columns in the dimension table Organize attributes as hierarchies Attribute hierarchies contain one All level and one leaf level User hierarchies are multi-level combinations of attributes Can be placed in display folders The important message to convey here is that dimensions let users “slice and dice” data from analytical perspectives. It may be useful to contrast dimensions in SSAS 2000 versus 2005/2008. Dimensions in SSAS 2000 are hierarchy-based, whereas dimensions in SSAS 2005/2008 are attribute-based. This significant architectural change provides rich dimension structure. Explain how these attributes will map to the columns contained in the dimension table and that these attributes form hierarchies in their own right (albeit a simple hierarchy with an All level and a single level beneath that containing the distinct members for the attribute). Multi-level hierarchies can be defined based on attributes where a one-to-many relationship exists between them (for example, Year to Quarter). Emphasize that the attribute nature of dimensions is what gives the UDM its relational flavor.

17 Dimension Types and Characteristics
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Dimension Types and Characteristics Time (Server time dimension) Dimension membership is based on a time range, not a dimension table Useful when basing the cube on an OLTP database Special: Account, Currency Manages aggregation behavior across accounts over time Adds currency conversion capabilities to a cube Characteristics Parent-child Write-enabled Linked Keep the explanation of this slide brief. There is a lot to discuss about dimension types and design, but this information is out of scope for this course. Examples of these types will be useful: Time: SSAS can generate a time dimension based on a date range. This is not a recommended practice for BI solutions based on a data warehouse (which should have its own date dimension). Special: Mention briefly that common scenarios of general ledger and currency exchange can be managed by SSAS using wizards. Characteristics: Parent-child: The dimensions are implemented as an attribute type within a dimension. Organization charts, general ledger, and bill of materials are all common examples of parent-child. Remind students that the number of levels in these attributes is determined by the data itself and not the structure of the dimension table. Write-enabled: Enables users to modify the contents of the dimension and see the immediate effect of changes on the hierarchies of the dimension—great for what-if analysis. Linked: Dimensions can be based on a dimension that is stored in another SSAS database.

18 Dimension Relationships
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Dimension Relationships Define interaction between dimensions and measure groups Relationship types Regular Reference Fact (Degenerate) Many-to-many Data mining The UDM can present data from multiple fact tables. Dimension relationships (where such relationships exist) need to be defined with the measure groups: Regular: A regular dimension relationship between a cube dimension and a measure group exists when the key column for the dimension is joined directly to the fact table. The majority of dimensions will have a regular relationship. Reference: A reference dimension relationship between a cube dimension and a measure group exists when the key column for the dimension is joined indirectly to the fact table by a key in another dimension table. For example, a Geography dimension may be used by both the Sales People and Customer dimensions. The star schema will appear snowflaked to support this arrangement. Fact: Ralph Kimball refers to this type as a degenerate dimension. Fact dimensions are built directly off a column in a fact table. Examples include tracking or order numbers that are stored directly in the fact table. Many-to-many: This relationship allows the joining of a single fact to multiple dimension members. For example, a customer may make multiple sales and assign multiple reasons to each of those sales. Many-to-many dimensions expand the dimensional model beyond the classic star schema and support complex analytics when dimensions are not directly related to a fact table. Data mining: Provide an example such as using a clustering algorithm to segment customers, then using the mining model to query a cube.

19 Working with Dimensions
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Working with Dimensions DEMO Refer to the demonstration notes. Mar-2008 Microsoft Developer & Platform Evangelism

20 Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Calculations Expressions evaluated at query time for values that cannot be stored in fact table Types of calculations Calculated members Named sets Scoped assignments Calculations are defined using MDX MDX = MultiDimensional EXpressions Begin by discussing how calculations in Excel work: You have data in cells A1 and B1, then in cell C1, you create a function such as SUM(A1:B1). Although this is a simplistic example, it makes it easier for most students to understand how calculations in a cube work. The cube contains base data, and calculations are on this data. Of course, an Excel spreadsheet is usually more simplistic than a multidimensional database! Calculations are written using MDX, which is a topic outside the scope of this course. Do not attempt to make this an MDX lesson. Time permitting, a demonstration of an MDX query may be appropriate in SQL Server Management Studio. The upcoming lab exercise has the students copy and paste expressions into their project. Provide examples of each calculation type: Calculated members can be created on any dimension, including the Measures dimension. Calculated members created on the Measures dimension are called calculated measures. A sales-variance calculation can be based on the Sales Amount and Quota Sales measures. Named sets: A named set is created for reuse in MDX queries (for example, the Top 10 Performing Sales People). Scoped assignments: You can define subcube expressions that allow the application of a formula to an isolated part of the cube space. For example, you can apply a different tax-rate calculation to sales made in a particular state. Mar-2008

21 Programming SSAS SSAS OLAP Programmability
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Programming SSAS SSAS OLAP Programmability SSAS Application Programming Interfaces Mar-2008

22 SSAS OLAP Programmability
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model SSAS OLAP Programmability C++ App VB App .NET App Any App OLE DB ADO ADOMD.NET AMO Any Platform, Any Device WAN XMLA Over TCP/IP XMLA Over HTTP Analysis Server OLAP Data Mining Server ADOMD.NET Data Mining Interfaces .NET Stored Procedures Microsoft Algorithms Third-Party Algorithms Use this slide to discuss the programming opportunities with SSAS OLAP. Note that this slide repeats in the data mining module (Module 08), so focus only on OLAP for now. The next slides discuss AMO, ADOMD.NET, Server ADOMD, and .NET stored procedures.

23 SSAS Application Programming Interfaces
Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model SSAS Application Programming Interfaces AMO (Analysis Management Objects) .NET based and suitable for managed code programs Develop management applications to administer database objects, security, processing, etc. ADOMD.NET Extension of ADO.NET Connect to SSAS databases to retrieve and manipulate data Server ADOMD.NET Extend MDX with .NET stored procedures AMO: The Analysis Management Objects (AMO) library provides a complete set of Microsoft .NET Framework objects for use in managed applications, such as those developed in Visual Basic .NET or C#. With the AMO library, you can manage SSAS 2005 objects, including servers, databases, data source views, cubes, dimensions, mining models, and roles. You can also administer security, process cubes and mining models, and so on. Object Explorer in SQL Server Management Studio is built on top of AMO. ADOMD.NET: (Client) ADOMD.NET is modeled after ADO.NET and ADOMD and has been designed mainly to send MDX queries to the server. Client applications can use ADOMD.NET to connect to multidimensional data sources for retrieving and manipulating data and metadata. Server ADOMD.NET: Server ADOMD.NET is designed to extend MDX itself. It works with objects such as sets, tuples, members, and so on. SSAS 2005 includes true server-side stored procedure support. Mar-2008

24 Microsoft BI Voyage Business Intelligence Fundamentals: Unified Dimensional Model Resources SQL Server 2008 Books Online – Analysis Services, technet.microsoft.com/en-us/library/ bb522607(SQL.100).aspx Links to technical resources, case studies, news, and reviews Mosha Pasumansky is the development lead for the Microsoft Analysis Services engine Mar-2008


Download ppt "Business Intelligence Fundamentals: Unified Dimensional Model"

Similar presentations


Ads by Google