Week 02.pdf
0 likes14 views
The document discusses multidimensional data representation and manipulation, focusing on data cube concepts relevant to business analysts, particularly in analyzing employee turnover and compensation factors. It covers data cube structure, dimensions, measures, and operational properties, including the use of Microsoft Multidimensional Expressions (MDX) for querying. Additionally, it contrasts SQL and MDX in terms of their querying capabilities and illustrates examples using a data cube model.
Week 02.pdf
- 1. Multidimensional data representation and manipulation Data Cube Concepts
- 2. 2
- 3. 3
- 4. 5
- 5. 6
- 6. 7
- 7. 8 Business Analyst Perspective Employee Turnover Management Location Marketplace Compensation Factor / qualitative variable Outcome Variable / quantitative variable - Retention focus - Flexibility options - City size - Transportation options to work - Demand level - Salary escalation - Health care - Pension - Base salary levels
- 8. 9 Data Cube Basics • Business analyst model – Factors or influencing variables of interest – Quantitative variables – Multidimensional arrangement • Terminology – Dimension: subject label for a row or column Can have more than 2 or 3 dimensions dimension may be city size or type of health plan offered – Member: value of dimension – Measure: quantitative variables/data stored in cells can have more than one measure in a cell
- 9. 10 Sales Data Cube Example
- 10. 11 Notes on Dimensions and Measures • Hierarchies: Member can have sub members (more detail) Location: country, region, state, zip code • Sparsity: Many cells are typically empty when dimensions are related May not sell all products in all regions Major problem with storing data cubes: compression of unused space • Measures: Derived measure: Common: unit sales * unit volume; sales per transaction Data cube engine must compute efficiently Multiple measures in cells
- 11. 12 Measure Aggregation Properties • “Aggregate Property” indicates allowable summary operations for measures • Additive – Summarized by addition across all dimensions such as sales, cost, profit – Sales can be summed across product, time, customer, … • Semi-Additive – Summarized by addition in some but not all dimensions such as time – Periodic measurements such as account balances and inventory levels – Account balance can be summed across customer branch – Account balance cannot be summed across time because balance is just a point in time measurement • Non-Additive – Cannot be summarized by addition through any dimension – Historical facts such as unit price for a sale
- 12. 13 Measure Aggregation Example • Dimensions – Course: course id, degree, department, and college – Student: student id, major, department, and college – Time: semester, academic year, academic decade • Measures: – Credit hours – Grade – Unit tuition (cost per credit hour) – Tuition (unit tution * credit hours) • Aggregation properties for measures: ? Credit hours: additive across all dimensions Grade: non additive but averageable such as grade point average Unit tuition: non additive and averageable but probably not useful as an average Tuition: additive
- 13. 14 Sales Data Cube Example
- 14. 15
- 15. 16
- 16. 17
- 17. 18
- 18. 19
- 19. 20
- 20. 21
- 21. 22
- 22. 23
- 23. 24 Microsoft Multidimensional Expressions (MDX) Language History 1997 MS OLAP specification 1998 MS OLAP services release 2001 mdXML specification by XMLA Council MS 2005 MDX revision • Defacto standard developed by Microsoft and later by the XMLA (XML for Analysis) Council – A web standards group
- 24. 25 MDX Usage • Foundation for Microsoft products and open source analytics software • SQL Server Analysis Services and Excel Pivot Tables • Hyperion, IBM, SAP (Systems Applications and Products), and other vendors • Foundation for open source projects: JPivot, Pivot4J, and Pentaho Business Analytics Platform
- 26. 27 Attribute Hierarchy and Members Line attribute member Vendor attribute member Product attribute member Line Vendor Product Product attribute hierarchy
- 27. 28 Steel Wheels Cube Display
- 28. 29 MDX Terminology Notes • Tuple – Cell identifier – One member from each dimension • Axis: dimension selected in a query (source cube cells) • Slicer: combination of dimension members (result cube cells) • Querying Multidimensional Data with MDX – https://docs.microsoft.com/en-us/analysis- services/multidimensional-models/mdx/querying- multidimensional-data-with-mdx?view=sql-analysis- services-2022
- 29. 30 SQL Versus MDX • Table result for SQL SELECT statement • Data cube result for MDX SELECT statement • Different mathematical approaches for manipulating tables (e.g. relational algebra) and data cubes (e.g. matrix algebra)
- 30. 31 Comparison of Clauses Language Clause SQL MDX SELECT List of columns List of axis dimensions (source cube cells) FROM List of tables Cube name WHERE Conditions restricting rows Restriction to a combination of dimension members (result cube cells)
- 31. 32 MDX query example : https://dba.stackexchange.com/questions/138311/good-example-of-mdx-vs-sql-for-analytical-queries
- 32. 33 Example MDX Statement and Result - Dimensions in the WHERE clause must be different than the SELECT clause - WHERE condition is known as a slicer condition
- 33. 34 CrossJoint Operation • Combines multiple dimensions or measures on a single axis
- 35. 36 Pivot Table • Powerful interface for data cubes • Convenient rearrangement of row and column headings • Expand or collapse dimensions
- 36. 37 Pivot4J • Allows cube representation similar to pivot table in Microsoft Excel • Works with Pentaho Business Analytics • Separate add-on • Graphical implementation of the MDX language
- 38. 39 Pivot Table with MDX Statement
- 39. 40 Pivot Table with CrossJoin