Case study · Pharmaceutical sector

A Data Lakehouse to unify and govern a pharma company's data

A cloud architecture that orders, governs and serves data to every therapeutic area and affiliate, with GxP governance as the foundation.

Layered Medallion architecture

Data governance with Unity Catalog

GxP governance

Where it all started

El punto de partida

A pharmaceutical company with operations in several countries ran its analytics on a legacy visual analytics platform. It worked, but it weighed more every day: ingestion was manual, based on stored procedures and scripts, with no central repository, and with batch processes in fixed windows that were hard to scale. And in pharma, on top of that, every piece of data must be governable and auditable (GxP), something the legacy platform did not make easy.

Every new report was a fight against the plumbing of the data.

The challenge

Manual, tightly coupled ingestion, hard to maintain.

Data scattered, with no central repository or modern architecture.

Rigid batch processes: bottlenecks as volume grows.

Permissions, lineage and traceability with no centralized control, in a sector that demands data governance (GxP).

What we proposed

Build a Data Lakehouse on Databricks with a Medallion architecture: a single cloud platform to ingest, transform and govern all the data, with Unity Catalog and GxP governance, ready to feed the reporting of every therapeutic area and affiliate, and to exploit RWD/RWE and AI.

The shift

On the outside, the same reports as always. On the inside, a data architecture rebuilt from top to bottom.

Before

Manual loads with stored procedures

Data scattered across ERP, CRM and SharePoint

Rigid batch processes in fixed windows

No centralized governance or traceability

Now

Automated ingestion from every source

A single, governed data layer

Elastic compute that scales with demand

Lineage, versioning and role-based access, with GxP governance

sources integrated (ERP, CRM, SharePoint)

Medallion layers

of data governed

1 %

The value generated

Beyond architecture

A single data layer

All business data consolidated: one single truth for CommEx and every area.

Governance and lineage

With Unity Catalog, every piece of data has an owner, permissions and traceability from origin to destination.

Ready to audit (GxP)

Encryption, access control and versioning (time travel) that support demands like GxP, GDPR or HIPAA.

Scale without rewriting

New sources, countries and therapeutic areas on the same architecture.

Cost under control

Elastic compute per use, with no fixed infrastructure to sustain.

A foundation for AI and RWD

The governed data is left ready for AI and for exploiting Real World Data (RWD/RWE).

What the client says

The data team went from maintaining manual loads and batch processes to working on a single governed platform. Today the data from its ERP, its CRM and its documents lives in one place, catalogued and traceable, and every new source or line of business is added on the same architecture, with nothing to rebuild.

Information Systems Management

International pharmaceutical company

F&Q

Frequently asked questions

Why Databricks and a Medallion architecture?

Databricks, on Apache Spark, offers elastic cloud compute: it processes large volumes and scales on demand, without the bottlenecks of traditional batch processes. The Medallion architecture organizes data in layers: Bronze (raw), Silver (consolidated) and Gold (modeled), so every report is fed by clean, traceable data.

How do you govern and protect sensitive, regulated data?

With Unity Catalog as the governance layer: a centralized catalog, role-based access control and full lineage of every transformation. Data travels encrypted at rest and in transit, and time-travel versioning makes it possible to audit and roll back changes. It is the foundation for sustaining GxP governance and demands like GDPR or HIPAA, common in the pharmaceutical sector, and for exploiting RWD/RWE with guarantees.

Can you migrate without stopping analytics operations?

Yes. We separate development and production environments and deploy with CI/CD: every change is validated before reaching production, without cutting off the reports the business is already using. And because the whole architecture is versioned in repositories, in the event of an issue it is redeployed in a controlled way, protecting service continuity.

What does it take to get started?

Your data sources (an ERP such as SAP, a CRM, SharePoint or others) and a cloud environment. From there we work in phases: we build the architecture, migrate a first line of business and put it into production before scaling to the rest. Value arrives early and risk stays contained.