• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
Visual Data Vault
 

Visual Data Vault

on

  • 396 views

My presentation on the Visual Data Vault modeling language, presented during WWDVC 2014 in St. Albans, VT, USA. ...

My presentation on the Visual Data Vault modeling language, presented during WWDVC 2014 in St. Albans, VT, USA.

To download the Visio stencils, check out
http://www.doerffler.com/know-how/data-vault/visual-data-vault/
and http://www.visualdatavault.com

Statistics

Views

Total Views
396
Views on SlideShare
248
Embed Views
148

Actions

Likes
0
Downloads
1
Comments
0

2 Embeds 148

http://www.doerffler.com 147
https://www.linkedin.com 1

Accessibility

Upload Details

Uploaded via as Microsoft PowerPoint

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Visual Data Vault Visual Data Vault Presentation Transcript

    • VISUAL DATA VAULT [MODELING LANGUAGE] MichaelOlschimke World-Wide DataVault Consortium, St.Albans,Vermont
    • Introduction Goals Basic Entities Query AssistantTables ReferenceTables BusinessVault Remarks AGENDA March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 2
    • INTRODUCTION Visual DataVault [Modeling Language] March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 3
    • • Mid-size consulting firm in Germany • Consulting, training, implementation • Focus on BI • Also: relational databases, mainframe, software development • Industries: • Automotive • Banking • Consumer • Pharmaceutical • Telecommunications • Insurance • Partners: INTRODUCTION (1/2) March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 4
    • • BIConsultant Dörffler + Partner GmbH • Specialized on DataVault, data mining, CRM, ETL, project management • DataVault 2.0 Certified Individual • Sectors: automotive, commerce, public, non-profits • Academic research on neural networks, text classification, information retrieval • Located in Germany INTRODUCTION (2/2) March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 5
    • GOALS Visual DataVault [Modeling Language] March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 6
    • Visually express DataVault models Generate DDL from DataVault models Microsoft Office support GOALS March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 7
    • 3. BASIC ENTITIES Visual DataVault [Modeling Language] March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 8
    • • A list of business keys • Business keys are attached to hub • Composite key is modeled by adding multiple business keys to hub • Business keys might have data types 3.1 HUBS CustomerCustomer Country Customer No. March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 9 Customer Country Customer No. Customer Country: varchar(2) Customer No.: integer
    • • Smart Keys are keys with some logical structure • Not a composite key • Do not model check sums • Do not model smart key if format is unclear or multiple format definitions are possible • Possible to integrate in composite key • Composite key might consist of multiple smart keys 3.1.2 SMART KEYS Vehicle Vehicle Identification Number Vehicle Descriptor Section World Manufacturer Identifier Vehicle Identifier Section March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 10 Vehicle Vehicle Identification Number Vehicle Descriptor Section World Manufacturer Identifier Vehicle Identifier Section Brand Vehicle Vehicle Identification Number Vehicle Descriptor Section World Manufacturer Identifier Vehicle Identifier Section Vehicle Bar Code Stock Number Parking Lot Number
    • • Links connect hubs • Relationships or transactions • Read: „Stock used by StockTrade“ • Check comments inVisio stencil • Link reference might be overwritten (add name to connector) • Important for multiple references of the same hub in one link • Possible to add attributes to links (e.g., degenerated fields) 3.2 LINKS Stock TradeStock Account Customer Account March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 11 Stock TradeStock Account Customer Account Diverted Flight Airport Source Airport Destination Airport Diverted Flight Airport Diversion Number Source Airport Destination Airport
    • • Special form of link • Data cannot legally change • Notice the annotation in the icon • Transactional satellites are discussed later 3.2.1TRANSACTIONAL LINKS T Sales T SalesProduct Customer T Sales Information Sales Status March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 12
    • • Link-to-Link structures can be modeled as well • However: not recommended because of load dependencies • Load dependencies complicate the automated loading 3.2.2 LINK-TO-LINK Supplier Sales Person Territory Product Product Sales Person Territory March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 13
    • • Satellites store descriptive data • Usually historized • Data is stored in attributes • Attached to hubs or links 3.3 SATELLITES (1/2) Shipping AddressShipping Address City Address Line 2 State Address Line 1 Zip Code Shipping Address City Address Line 2 State Address Line 1 Zip Code Customer March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 14
    • • Links and hubs might have multiple satellites • Small bug in MSVisio stencil 3.3 SATELLITES (2/2) Audit Information Quantities Stock Trade Turbulence March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 15
    • • Usually data comes from multiple sources • Record tracking satellites track the availability of keys and associations in source systems • Special satellite variant • Normalized or de-normalized version is not indicated (physical features are not covered by the modeling language) 3.3.1 RECORDTRACKING SATELLITES (1/2) Customer Customers from CRM Customers from Invoicing Customers from Web Shop R Customer Tracking Satellite Customer Customers from CRM Customers from Invoicing Customers from Web Shop R Customer Tracking Satellite March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 16
    • • Link version of record tracking satellite • Follows the hub version (record tracking satellite can be added to hub or link) 3.3.1 RECORDTRACKING SATELLITES (2/2) Sale Sale Information from CRM Sale Information from Analytics Sale Information from Web Shop Sale Sale Information from CRM Sale Information from Analytics Sale Information from Web Shop R Sale Tracking Satellite Turbulence Fasten Your Seatbelt March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 17
    • • Attached to hub or link • Follows general satellite structure • There is always a Status attribute 3.3.2 STATUSTRACKING SATELLITES Customer Customer Status StatusCustomer Customer Status StatusCustomer Customer Status Status March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 18
    • • Alternative to transactional links • Transactional satellites are attached to transactional links • They store no history • Attributes are added to the satellite structure • Introduced to allow automated generation of DDL from such models 3.3.3TRANSACTIONAL SATELLITES Product Customer T Sales TransactionProduct Customer T Sales Transaction T Sales Transaction Data Quantity Item PriceTotal Price March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 19 Product Customer T Sales Transaction T Sales Transaction Data Quantity Item PriceTotal Price
    • 4. QUERY ASSISTANT TABLES Visual DataVault [Modeling Language] March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 20
    • • PIT table spans the satellites of one hub or link • Implemented as a ribbon that is attached to the hub or link symbol • All satellites are affected by the PIT 4.1 POINT-IN-TIME (PIT)TABLES Contact March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 21 Contact CRM Leads Newsletter ContactsArticle Reviewers
    • • Bridges improve join performance between hubs and links • Hub or link is “used by” bridge 4.2 BRIDGES (1/2) Bridge Product Parts Customer Bill of Material T Sale Bridge Marketplace Shop March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 22
    • • Also possible to overwrite the reference name 4.2 BRIDGES (2/2) Product Customer Bridge Lead March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 23
    • 5. REFERENCE TABLES Visual DataVault [Modeling Language] March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 24
    • • Reference tables are lookup tables that store descriptive data • Have at least one business key • Have multiple attributes • Business key might be a smart key • Business key might be composite key • No history (flat structure) 5.1 NO-HISTORY REFERENCETABLES ColorColor Color Code Short Description Long Description March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 25 Color Color Code Short Description Long Description Color Detailed Color Identifier Short Description Long Description Color Code Main Color Identifier Color Detailed Color Identifier Short Description Long Description Color Code Main Color Identifier Product
    • • Similar to no-history reference table • Has business key in table • Satellite stores attributes with history-tracking • Satellite follows standard rules for satellites 5.2 HISTORY-BASED REFERENCETABLES Category Code Short Description Long Description Category Descriptions Category Code Short Description Long Description Category Descriptions Category Code Short Description Long Description Category Descriptions March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 26
    • • Master code table for commonly used codes and their descriptions • Reference table contains two business keys (Code & Group) • History-based Satellite for the descriptive attributes 5.3 CODE AND DESCRIPTIONS Master Code Table Code Short Description Long Description Master Code Attributes Group Master Code Table Code Short Description Long Description Master Code Attributes Group Master Code Table Code Short Description Long Description Master Code Attributes Group Master Code Table Code Short Description Long Description Master Code Attributes Group March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 27
    • 6. BUSINESS VAULT Visual DataVault [Modeling Language] March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 28
    • • Computed satellites describe a hub or link with computed descriptive attributes • Added to the hub or link in the same way as standard satellites • Computed attributes are added to the satellite • Might contain non-computed attributes (e.g. attributes that are duplicated from another satellite for convenience) 6.1 COMPUTED SATELLITES Invoice Totals Sales Invoice Total Grant Total Tax Rate March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 29 Invoice Totals Sales Invoice Total Grant Total Tax Rate Invoice Totals Sales Invoice Total Grant Total Tax Rate Invoice Totals Sales Invoice Total Grant Total Invoice Totals Sales Invoice Total Grant Total Tax Rate
    • • Concept is similar to a bridge • Changes the grain of a link • Aggregates values and adds them as computed attributes to the link 6.2 COMPUTED AGGREGATE LINKS Sales per Shop and Customer SaleCustomer Product Shop Total Sales Sales per Shop and Customer SaleCustomer Product Shop March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 30 Sales per Shop and Customer SaleCustomer Product Shop Total Sales
    • • These links are not available in source systems • Added artificially to the Data Vault for exploration purposes • Connects hubs that are not directly connected in source systems • Basket Analysis 6.3 EXPLORATION LINKS Customer Store Product T Sale March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 31 Offerings Customer Store Product T Sale
    • • BusinessVault tables have no prescribed format • Have business keys and attributes • Might have computed attributes • Might have computed satellites attached • Can be added to the Raw Data Vault by ordinary links that reference the primary key of the BusinessVault table 6.4 BUSINESSVAULTTABLES Customer First Name Last Name Customer Number Customer First Name Last Name Customer Number Customer First Name Last Name Customer Number City Address 1 Zip Code Computed Customer Attributes Life-Time Value of Customer Birth Date Customer First Name Last Name Customer Number City Address 1 Zip Code Computed Customer Attributes Life-Time Value of Customer Birth Date Customer Last Name First Name Customer Number City Address 1 Zip Code SalesProduct Computed Customer Attributes Life-Time Value of Customer Birth Date March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 32
    • SOME REMARKS Visual DataVault [Modeling Language] March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 33
    •  Logical modeling, no physical features  VisioThemes are not supported (yet)  More features to come:  Inline attributes  Validation rules (prevent hub on hub, etc.)  What else?  Don’t copy fromVisio and paste intoWord or PowerPoint  Instead: export toWMF for better quality  Vendor support package available  Check out www.datavault.de for German assets on DataVault REMARKS March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 34
    • March 20, 2014 World-Wide Data Vault Consortium, St. Albans, Vermont 35 Give us Feedback http://tinyurl.com/doerffler-wwdvc Source: vasilijonline.com