Agile BI via Data Vault and Modelstorming

Preview:

DESCRIPTION

Audience: Business Intelligence Architects, Project Managers and Sponsors. This slideshow accompanies a video presentation of the same name, available at http://youtu.be/e0cHFdeGEeE.

Citation preview

Agile BI via Data Vault and Modelstorming

Daniel Upton

Business Intelligence Architect, Certified ScrumMaster

DecisionLab.Net

linkedIn.com / in / DanielUpton

The Business Intelligence Promise: Smarter, more fact-based decision-making as an everyday routine

How to maximize BI stakeholder satisfaction and accelerate time to completion…

1. Break down stakeholder requirements gathering sessions to result in concise Data Stories 2. Complete a potentially shippable increment (PSI) in all, or nearly all, Sprints3. With each PSI, be ready and able to re-prioritize after every Sprint.

Business Intelligence Waterfall

Data Model

Profile

Stage Data (ETL)

Integrate Staged Data (ETL)

Requirements

Build Semantic Layer

Load to Star Schema (ETL)

Build Dashboard

Train Users

Quality Assurance On Everything

Release

Maintain

Entity Relational Model for Transactions

Dimensional Model / Presentation Layer

Business Intelligence

Risk

Report A Still

Doesn’t Match Report B

Business Intelligence Waterfall

Data Model

Profile

Stage Data (ETL)

Integrate Staged Data (ETL)

Requirements

Build Semantic Layer

Load to Star Schema (ETL)

Build Dashboard

Train Users

Quality Assurance On Everything

Release

Maintain

Scrum Development Lifecycle

Traditional Waterfall Principle:Big Design Up Front (BDUF) -- An end-to-end design on which to build an entire solution.

Lean Principle:Just Enough Design Up Front (JEDUF) -- A design increment with just enough design features on which to build and deliver an incremental solution that satisfies a required feature’s Acceptance Criteria.

Agile Principle: Maximize the amount of work not done.

BI’s most time intensive phases

Data Vault Tables: Hubs, Satellites and Links

The most time intensive phases

…accelerated by Data Vault

Faster, but not yet Agile-Fast

Agile ModelStorming

Agile ModelStorming

Agile ModelStorming

Agile ModelStorming

Review time intensive phases

Breakdown each step into ‘Data Story’ size

Assumption: Same overall amount of work

Accelerate delivery of data stories. Set up multiple team-tracks. Be ready for changing priorities.

The most time intensive phases

…accelerated with Agile DW Design

Data Vault already accelerated the Waterfall

Agile DW design and Data Vault accelerate each other

Agile Data Vault within Sprint Cycles

Result: Maximize stakeholder satisfaction and accelerate time to completion. How? 1. Break down stakeholder requirements gathering involvement into concise Data Stories 2. Complete a potentially shippable increment (PSI) in all, or nearly all, Sprints3. Be Agile: With a PSI, be ready and able to re-prioritize after every Sprint.

Recommended Technical Reading…

For an introduction to Data Vault…

“Data Vault: Data Warehouse Design Goes Agile”

www.slideshare.net/DanielUpton

The Business Intelligence Promise:Smarter, more fact-based decision-making as an everyday routine

How to maximize BI stakeholder satisfaction and accelerate time to completion…

1. Break down stakeholder requirements gathering sessions to result in concise Data Stories 2. Complete a potentially shippable increment (PSI) in all, or nearly all, Sprints3. With each PSI, be ready and able to re-prioritize after every Sprint.

LinkedIn.com / In / DanielUpton

Recommended