Skip to content
Datalloy

Service

Data Warehousing

A warehouse is where an organisation agrees what a customer, an order and a margin actually mean. Datalloy designs dimensional models that answer business questions quickly and keep the history needed to compare one period with another.
The data warehousing stage covers curated data, staging, dimensions, facts, data marts.

Where this sits

  1. 01

    Curated Data

    Cleansed, standardised source data

  2. 02

    Staging

    Load-ready structures and keys

  3. 03

    Dimensions

    Conformed, versioned, surrogate-keyed

  4. 04

    Facts

    Defined grain, additive measures

  5. 05

    Data Marts

    Subject-area models for reporting

Scope

What data warehousing covers

Dimensional warehouses and data marts that hold history, resolve conflicting definitions and stay fast as volumes grow.

  • Dimensional modelling

    Star schemas built around business processes: conformed dimensions, a clearly defined grain and fact tables that stay performant as they grow.

  • History and change

    Slowly Changing Dimensions, so last year's report still reflects last year's structure rather than today's organisational chart.

  • Data marts

    Subject-area marts for finance, operations, sales or supply chain, built on shared dimensions instead of separate, conflicting extracts.

  • Performance and scale

    Partitioning, indexing and load strategies designed around the query patterns the business actually runs.

  • Enterprise data warehouse
  • Star schema design
  • Fact tables
  • Dimension tables
  • Slowly Changing Dimensions
  • Data marts
  • Historical data

Signals

When this is usually needed

  • Two departments report different totals for the same measure
  • Historical comparison is impossible because the source overwrites records
  • Every new report needs a new extract built from scratch
  • Reporting queries are slowing down operational systems
  • Business definitions live in spreadsheets and in individual heads

Deliverables

What you get

  • Target data warehouse architecture and layer design
  • Bus matrix mapping business processes to conformed dimensions
  • Fact and dimension specifications with a defined grain
  • Slowly Changing Dimension handling per attribute
  • Incremental load and historisation logic
  • Data dictionary and model documentation

Next

Related services

The stages either side of this one — most platforms need more than a single layer.

  • Data Engineering

    Ingestion, transformation and orchestration built as engineered software — version controlled, parameterised, monitored and repeatable.

  • Business Intelligence

    Power BI semantic models, governed measures and dashboards designed around the questions the business actually asks.

Need data warehousing help?

Tell us where the data currently breaks down and we will tell you what we would do about it.

Start a ConversationExplore Our Services