Data Warehouse Build

A charter, source inventory and bus matrix, a layered architecture with standards and environments, star schemas with honest grain and history, and the reviews, migration and cutover work that make a warehouse trusted.

Free · Mac (Apple silicon) and Windows · v1.0.5 · Other platforms

Then open the Marketplace in the app, search for Data Warehouse Build, and click Install. It goes into your own files, and works offline from then on.

What it installs.

1

Pillar

1

Area

50

Projects

205

Starter tasks

📊 Data & Analytics

Data Warehouse Build

Designing and building a central data warehouse that combines data from many systems into reliable, well-modelled tables for analysis.

  1. Warehouse charter tied to twenty business questions
  2. Source system inventory and access map
  3. Enterprise bus matrix of processes and dimensions
  4. Layered architecture from raw to marts
  5. Naming conventions and modelling standards guide
  6. Development, test and production warehouse environments
  7. Warehouse role and schema permission design
  8. Warehouse schema and infrastructure as code
  9. Weekly warehouse build backlog review
  10. Schema change review and release train
  11. Monthly reconciliation of core facts to source
  12. New source onboarding checklist
  13. Monthly stakeholder demo of new warehouse tables
  14. Quarterly model health review
  15. Warehouse operations runbook
  16. Architecture decision record log
  17. Table deprecation and retirement process
  18. Dimensional modelling fundamentals
  19. Declaring the grain of every fact table
  20. Slowly changing dimensions in practice
  21. Transaction, snapshot and accumulating fact tables
  22. Surrogate keys and late-arriving records
  23. Reading query plans for warehouse workloads
  24. Normalised core versus dimensional core decision
  25. Conformed customer dimension
  26. Shared date and fiscal calendar dimension
  27. Historical backfill strategy
  28. Partitioning and clustering plan for large tables
  29. Workload isolation for loading, modelling and BI
  30. Star schema versus wide tables for BI
  31. Handling deletes and corrections from source systems
  32. Currency and unit conversion reference tables
  33. First subject area mart launch
  34. Parallel run and report sign-off before migration
  35. Legacy warehouse cutover weekend
  36. Warehouse architecture review with the design authority
  37. Ninety-day post-launch retrospective
  38. Legacy warehouse decommissioning
  39. First data engineer building a startup warehouse
  40. Consultancy build handover to the in-house team
  41. Multi-business-unit warehouse federation model
  42. Merging warehouses after an acquisition
  43. Warehouse build in a regulated industry
  44. Rescuing a stalled or distrusted warehouse
  45. Onboarding an engineer to the warehouse codebase
  46. Data vault core layer pilot
  47. Open table format lakehouse layer design
  48. Bitemporal history for point-in-time reporting
  49. Multi-region warehouse layout for data residency
  50. Warehouse disaster recovery and restore drill

Where it starts.

Each project arrives with its purpose, milestones and first tasks written out. Here is how the first three read.

Warehouse charter tied to twenty business questions

This is a team plan, best installed into a shared workspace so everyone with a role in it can see and own their tasks. Warehouses that start from a list of source systems end up as expensive copies of those systems that nobody queries.

First tasks

  • Book thirty-minute interviews with five team leads who rely on reports
  • Ask each lead for the three questions they cannot answer today
  • Rank the questions by value and group them into subject areas
  • Ask the agent to draft the one-page charter from your interview notes
  • Revisit the charter questions with the sponsor and retire answered ones @recurring(yearly)

Source system inventory and access map

Every warehouse build stalls at some point on a system nobody knew about or a database nobody can get credentials for. An inventory of each source, its owner, its volume, how data can be extracted and what approvals are needed surfaces those blockers in week two rather than month four.

First tasks

  • List every system named in the charter interviews in a shared sheet
  • Add owner, technical contact and extraction method for each source
  • Estimate row counts and daily change volume for the ten largest tables
  • Raise access requests for every source and note the approver
  • Flag sources holding personal data for the privacy lead

Enterprise bus matrix of processes and dimensions

A bus matrix puts business processes such as orders, shipments and invoices down the side and shared dimensions such as customer, product and date across the top. Drawing it before modelling anything shows which dimensions must be conformed, which process to build first and where two teams are about to model the same thing twice.

First tasks

  • List the business events behind the top ten charter questions
  • Mark which shared dimensions each event needs
  • Circle the dimensions used by three or more processes as conformed
  • Walk two business owners through the matrix and fix the names

Make it yours.

Install Data Warehouse Build, then tell Aurum the date you're aiming for. It checks whether the plan fits the weeks you actually have.

Free · Mac (Apple silicon) and Windows · v1.0.5 · Other platforms

Blueprints are general information to help you organise your own work. They are not medical, financial, legal or other professional advice. See the Terms.