This repository contains anonymised Power Query (M) and SQL logic used to automate Completeness & Accuracy (C&A) testing for G-SIB (Global Systemically Important Banks) and high-risk financial audits.
During large-scale financial audits, manual reconciliation of transactional data across disparate systems (GCP, BigQuery, Tableau) is prone to human error and logic gaps, especially during mid-period system migrations.
I developed this logic to perform a Full Outer Join between front-end reporting layers and back-end database sources. This ensures:
- Completeness: Identifying "Orphan" records that exist in the source but failed to reach the report.
- Accuracy: Validating field-level integrity (Account IDs, Exposure amounts, Run UUIDs) across 180,000+ records.
- Logic Consistency: Verifying that SQL-based alert configurations remain consistent after cloud migrations (e.g., GCP to BigQuery).
Below is a visual representation of the reconciliation join logic.
Click the link above to view the documented, copy-ready code used for this automation.
The logic in this repo demonstrates:
- Power Query (M): Nested joins and column expansion for multi-source reconciliation.
- SQL: Logic integrity checks and data profiling.
- Tools: Alteryx, BigQuery, Visual Studio Code.
Note on Confidentiality: All data, account identifiers, and client-specific parameters have been anonymised or replaced with mock values to maintain strict data privacy and professional ethics. These artefacts represent personal methodology and best practices.
Return to main profile: github.com/lbhatti-risk