One Source of Truth: Consolidating 50+ Sources on Azure

Decades of brand and country expansion had left Alshaya with 50+ disconnected data sources. Reconciliation was manual, definitions conflicted, and no function could trust a cross-enterprise number.

An Azure Data Lake on Databricks with a medallion-style progression — raw ingestion, conformed/cleansed, and curated marts — serving Merchandise (Food & Non-Food), Corporate, Finance, Customer, and Logistics.

Structured the ingestion and curation program with clear ownership and quality gates, then wired the curated layer into an Oracle ADW reporting platform with 60+ enterprise reports and self-service BI.

• Harmonizing conflicting definitions across brands and geographies.
• Ingesting both structured and unstructured data reliably.
• Delivering trust, not just a pipeline.

Consolidated 50+ sources into a single governed lake, established conformed definitions, and delivered near-real-time reporting with a 20% batch-processing improvement.

• A single source of truth adopted across five enterprise domains.
• 20% faster batch processing and fresher executive reporting.
• Elimination of manual cross-source reconciliation.

A data lake is only as valuable as the definitions layered on top of it. Governance and conformance are the real deliverables.

Streaming ingestion, a semantic layer over the curated marts, and ML-ready feature stores for forecasting.