Data systems for growing businesses
I turn scattered systems into one governed pipeline and a dashboard people actually trust.
Spreadsheets, a CRM, an e-commerce platform, a marketing tool. If your business runs on data spread across three places that don't agree with each other, I build the pipeline and the dashboard that fixes it.
What I do
Three related things, in order of how most engagements actually start.
Data pipeline builds
One reliable source of business data, instead of three spreadsheets that don't agree. I connect your databases, spreadsheets and SaaS tools into a single automated pipeline, so nobody has to manually combine exports or work out which version is correct. Built as a tested, versioned Bronze/Silver/Gold pipeline with automated data-quality checks at every stage.
Dashboards people actually trust
One dashboard everyone in the business actually trusts, with every number traceable back to a single source of truth instead of three competing spreadsheet exports. It's Metabase on top of governed, pre-aggregated Gold tables, so it stays fast even as the data grows.
Light app & product work, when the data need is really "capture this properly"
Sometimes the real blocker upstream of good data isn't the pipeline. It's that nothing captures the data cleanly in the first place, and I can build the small app that fixes it.
Real output, not a screenshot
The numbers here are real: a static export of Gold-layer tables from a pipeline built exactly like the ones I deliver for clients, run end-to-end on a public dataset instead of a mock-up.
Monthly revenue (R$)
On-time delivery rate
A real pipeline run on the public Olist e-commerce dataset (Jan 2017 to Aug 2018), not a mock-up. Numbers come straight out of this pipeline's Gold-layer tables. Read the full architecture write-up.
Is this you?
A few signs this is worth a conversation:
- • You're manually combining spreadsheet or CSV exports every week just to produce one report.
- • Your booking system, CRM, e-commerce platform and marketing tool don't agree on the same numbers.
- • You've already got dashboards, but nobody fully trusts what's on them.
- • Reporting depends on one person knowing how all the pieces fit together.
- • You want a proper data platform, not a full-time data engineer, at least not yet.
Case studies
One anonymized client engagement, and one fully public worked example showing exactly how the pipeline underneath is built.
From scattered spreadsheets and two disconnected systems to one governed pipeline
A small events-industry business had customer, booking, and email-marketing data spread across an operational database and a marketing platform, with no single trustworthy view of the business. I built a governed batch pipeline and a dashboard the team actually uses daily.
- Dashboard tabs
- 10
- Dashboard cards
- 90+
- SCD2 dimensions
- 8
- Automated tests
- 65+
How I build this: a public, fully-worked pipeline on the Olist e-commerce dataset
The exact same pipeline I build for real clients (medallion architecture, SCD2 versioning, data quality checks at every layer, Airflow orchestration), just run on a public dataset instead of confidential client data, so there's nothing to take on faith. It's the clearest way to see how I actually work.
- Orders processed
- ~100k
- SCD2 dimensions
- 3
- Fact tables
- 4
- Gold tables
- 4
About
I'm a solo data platform contractor helping growing businesses turn data that has outgrown ad hoc spreadsheets into a properly engineered platform, without needing to build an internal data engineering team. I design and build the pipeline, apply the governance (schema enforcement, versioned history, data quality checks) that makes the numbers trustworthy, and hand over a dashboard the business actually uses, not a demo that gets opened once and forgotten.
Occasionally the real blocker is upstream of any pipeline. The business needs a small, focused application to capture data cleanly in the first place, and I can build that too, when it's genuinely the right fix.
Get in touch
Tell me what's messy. What systems you're using, and what you're currently doing by hand to make the numbers agree. I can usually tell pretty quickly whether there's a sensible fix.
Prefer email? hello@theowhite.dev