This system is running live
Don't take
our word for it.
Try it.
We rebuilt a miniature version of this engagement and it's running right now. Click below — you are not looking at a screenshot, you are looking at the actual software.
▶ Open the live demo →Logistics · Client work · 11 weeks
Data Platform Modernisation
Lifting a fragile spreadsheet stack onto Postgres + dbt
- duration
- 11 weeks
- recalc time
- 45min → 2min
- months without failure
- 6
The problem
A regional logistics company ran their entire operational reporting on a single Excel workbook. 2,400 rows, 14 years of accumulated VBA macros, three people who knew how it worked, and one of those three was about to retire. Every Monday morning the workbook took 45 minutes to recalculate.
What we built
- **Postgres** as the new source of truth
- **dbt** to model the existing business logic from VBA into SQL transformations
- **Metabase** for the dashboard layer the operations team uses every day
- **Airflow** for the daily ingestion from the upstream systems
- **A migration that ran for 4 weeks in parallel** with the old workbook so the team could verify every number matched before switching over
What we deliberately did not do
- We did not migrate to a "modern data stack" with five vendors. Postgres + dbt + Metabase is enough for a company this size and costs ~$0/month to run.
- We did not rebuild the dashboards from scratch. We replicated the existing ones first, then iterated based on what the team actually wanted.
- We did not touch the upstream systems. Their API was stable and changing them was out of scope.
Outcome
- Monday morning recalculation: 45 minutes → 2 minutes
- Reporting reliability: from "the workbook crashes every other week" to zero failures in 6 months
- Knowledge spread across the team — not bottlenecked on the retiring engineer
Tech stack
PostgreSQL 16, dbt, Airflow, Metabase, Python ingestion scripts.