Skip to content
← All case studies
BI delivery and ERP-connected reporting Internal work sample

Governed Power BI models for operational and financial reporting

Delivered two source-controlled semantic models across 128 entities, giving operational and financial reporting a reviewable, governed foundation.

2 source-controlled
Semantic models
~35 tables
Financial star
33
DAX measures
plus 8 calculated columns
128
Entities served

Evidence inputs

  • Two Power BI semantic models over group data spanning 128 entities; one operational (PMO), one financial.
  • TMDL v3 / PBIP · DAX · Power Query (M)

Transformation

  • Authored eight calculated columns and thirty-three measures for reconciliation, variance and many-to-many bridging.
  • Structured the financial model as a conformed star of roughly thirty-five tables.
  • Kept both models under source control and verified edits against a written pre-flight checklist.

Controls & assurance

  • Used source-controlled TMDL so model changes are reviewable and diffable.
  • Required a pre-edit check to prevent load-breaking errors.

Output

  • Two governed semantic models — eight calculated columns and thirty-three measures over a roughly thirty-five-table financial star across 128 entities — delivered under a repeatable, source-controlled process.

Business value

  • Replaces drifting reports with a governed, reviewable foundation.
  1. Problem

    Operational and financial reports lacked governed models, making changes difficult to review and prone to drift.

  2. Approach

    Authored eight calculated columns and thirty-three measures for reconciliation, variance and many-to-many bridging.

  3. Outcome

    Two governed semantic models — eight calculated columns and thirty-three measures over a roughly thirty-five-table financial star across 128 entities — delivered under a repeatable, source-controlled process.

Business value

Replaces drifting reports with a governed, reviewable foundation.

Transformation route

  1. 01

    Authored eight calculated columns and thirty-three measures for reconciliation, variance and many-to-many bridging.

  2. 02

    Structured the financial model as a conformed star of roughly thirty-five tables.

  3. 03

    Kept both models under source control and verified edits against a written pre-flight checklist.

Decision log

  • Used source-controlled TMDL so model changes are reviewable and diffable.
  • Required a pre-edit check to prevent load-breaking errors.

This case study proves

  • DAX Data & Reporting Strong
  • TMDL semantic modelling Data & Reporting Strong

Full skill evidence →

What I learned

  • Source control and a pre-flight gate make semantic-model editing repeatable.

Future improvements

  • Complete the financial-star refactor and publish a reusable measure library.

← All case studies · Discuss this work →