Industry · Bizkaia
Industrial company
Data platform on top of the corporate ERP
The client
An industrial company in Bizkaia with several years of history accumulated in its ERP. By agreement with the client, we do not publish their name.
The challenge
Business-critical data was trapped in the ERP (Oracle/Prinex) and scattered across parallel spreadsheets. There was no single view: every report was rebuilt by hand, two people could give different figures for the same question, and nobody knew for certain which source was the right one. Querying the ERP directly was not an option either, because neither its data model nor its performance are designed for that: an analytical query can slow down the very system daily operations depend on. And the accumulated history brought what history always brings — formats that changed over the years, fields that mean different things depending on the period, and duplicates somebody has to decide how to resolve.
What we did
We designed an end-to-end pipeline: extraction from the ERP, an ETL that cleans and normalises, a PostgreSQL database designed for querying, an API and a frontend where the business consults the consolidated information and produces its reports without depending on anyone. The process runs unattended and records every load, so if a figure looks surprising you can trace where it came from and which extraction it belongs to. Separating the analytical database from the ERP has a further benefit: querying no longer competes with operations.
Technologies
Oracle · Python · ETL · PostgreSQL · FastAPI · Next.js · Azure
Result
A single source of truth in production: the same question gets the same answer, and the business consults the information and produces its reports without depending on somebody assembling a spreadsheet. When a figure stands out, it can be traced back to its origin instead of argued over.