<< VIEW FULL RESOURCE LIBRARY

Data Modeling Comparison: Tableau, Cognos & Power BI

January 16, 2020

Analytics Demystified, Cognos Analytics (v11), Data Preparation, IBM Cognos, Metadata Modeling, Microsoft Power BI, Tableau

Demo and Discussion of Approaches, Similarities and Differences

A well-constructed data model saves report authoring time and ensures data accuracy. The key to self-service BI, data modeling irons out the wrinkles created by data warehouse storage, diverse data sources and mismatched data categorization and puts it all in easily accessible, business friendly terminology. While Cognos, Tableau and Power BI all provide ways to model data, the options, methods and approaches for creating models in each differ significantly.

In this on-demand webinar, we take to the data modeling runway and strut our stuff. We peek behind the scenes of great looking visualizations and demo how modeling and transformations occur when using Tableau, Power BI and Cognos. We provide a clearer understanding of the capabilities available within the three tools and how they measure up.

READ THE ANSWERS TO THE QUESTIONS ASKED DURING THE WEBINAR

This event isn’t your off-the-rack presentation. It’s haute data couture, baby! We are partnered with all three of the market leaders and fluent across all the platforms. This is your chance to get an unbiased tool comparison.

▼ TECHNOLOGIES COVERED

Cognos, Microsoft Power BI, Tableau

▼ PRESENTER

Michael Weinhauer
Director of Training and Content
Senturus, Inc.

In addition to heading up the training at Senturus, Michael’s team is responsible for the development of the Senturus Analytics Connector, which lets Tableau and Power BI use Cognos as a data source. He has been designing, delivering and selling analytics solutions for over 20 years. Before Senturus, Michael held positions at Oracle, IBM and SAP, acquiring a wealth of hands-on, practical BI and big data experience. 

▼ PRESENTATION OUTLINE 1

  • BI design stages
    • Source—model and transform—visualize and analyze—share and collaborate
  • What is modeling and transformation?
    • Creating analysis-friendly data sets (virtual stars)
    • Joining, cleaning and otherwise “shaping” data
    • Data warehouse—single source of truth
    • Tools to prepare and model data
    • Roles are changing
  • Cognos Framework Manager client
    • Relational
    • Multidimensional
    • ODBC
    • JDBC
    • SAP
    • XML
    • Salesforce
    • Data sets
    • Other metadata sources
  • Cognos Framework Manager
    • Relational
    • Multidimensional
    • Cloud
    • SAP
    • Teradata
    • Salesforce
    • Data sets
    • Flat files
    • XLS
  • Cognos data modules
    • Existing data modules
    • Pre-configured database connections
    • Files (Excel, CSV)
    • Data sets
    • Packages
  • Cognos options diagram
  • Tableau
    • Relational
    • Multidimensional
    • ODBC
    • JDBC
    • SAP
    • Flat files
    • Statistical
    • Spatial
    • JSON
    • Python (TabPy)
    • PDF
    • Application specific
    • Live or extract
  • Tableau Prep
    • Relational
    • Multidimensional
    • ODBC
    • JDBC
    • SAP
    • Flat files
    • Statistical
    • PDF
    • Application specific
  • Power BI
    • Relational
    • Multidimensional
    • ODBC
    • JDBC
    • SAP
    • Flat files
    • Statistical
    • JSON
    • PDF
    • Python
    • Web URL
    • SaaS applications
  • Model & transform – Cognos Framework Manager
    • Multi-tier design
      • Namespaces and folders used to organize “query subjects”
      • Allows for multiple layers of abstraction and organization
      • Model dimensionally

▼ PRESENTATION OUTLINE 2

  • Set format, aggregations, data types
    • Resolve potentially complex design issues
    • Join ambiguity and “loops”
    • Recursive joins
    • Multiple logical hierarchies
    • Fact-to-fact (many-to-many) joins
    • Blind spot queries
    • Role-playing dimensions
  • Apply package and object-level security
  • Set governors
  • Set query timeouts
  • Allow/disallow cross-product joins
  • Allow/disallow outer joins
  • Set large text item limits
  • Publish packages for use by authors
    • Model & transform – Cognos data modules
      • Self-service web-based modeling
      • Create relationships
      • Filter data
      • Format
      • Split
      • Clean
      • Change data type
      • Create hierarchies
      • Define security*
      • Combine sources
      • Organize
      • “Intent” – AI/keyword search model suggestions
    • Model & transform – Tableau
      • Single pane (per data source)
      • View properties by selecting “describe”
      • Splitting
        • String fields can be split into multiple fields for easier analysis
        • Automatic or custom split options
          • Split based on a common separator
        • Aliasing
          • Roles – i.e., time (ship date/order date)
          • Binning (high/low sales)
        • Renaming
        • Data typing
        • Geographic roles
        • Calculations
        • Pivoting
        • Data interpreter
        • Hierarchies – basic
        • Time intelligence
  • Model & transform – Tableau Prep
    • Visual data preparation tool
    • AI/fuzzy logic driven
    • Tableau-centric – currently outputs only to flat file or extract
  • Model & transform – Power BI query editor layout
    • Ribbon (Home, Transform, etc.)
    • Queries pane
    • Query settings pane
    • Data preview of table
  • Model & transform – Power BI query editor layout
    • Query context menu
    • Table context menu
    • Column data type context menu
    • Column context menu
  • Model & transform – Power BI data view
    • Inspection
    • Exploration
    • Understanding
  • Model & transform – Power BI model view
    • Break out complex models into separate diagrams
    • View and modify properties
    • Set display folders to simplify navigation
  • Comparison of modeling & transformation capabilities
  • Summary
  • Lots of ways to get from point A to point B
  • You can probably get there with any of the tools
  • Question of, among other things:
    • Analytics culture (self-service vs. centralized)
    • Skillsets
    • Tool choice
    • Use cases/business requirements
    • Data source(s)