Data integration · BigQuery · Google Cloud
All your business data in one place you can trust.
Your numbers live in a route accounting system, a POS, a supplier portal, three spreadsheets and someone's inbox. We connect them, clean them and land them in BigQuery, so reports, apps and AI agents all read the same truth.
What we connect
Business systems
ERPs, route accounting and distribution systems, CRMs, payroll and accounting tools, through their APIs or scheduled exports.
Files people send
CSV and Excel files from email, SFTP or Cloud Storage, mapped by column name so a reordered or extra column doesn't break the load.
Portals with no API
Legacy reporting sites behind logins and SSO. We automate the export with a headless browser and schedule it.
Devices in the field
POS terminals, fiscal printers, card payments and GPS trackers, streaming into one backend.
Google Workspace
Gmail attachments, Drive folders and Sheets, read and written through the Google APIs.
Your own apps
Product databases (MongoDB, Postgres) synced to BigQuery for analytics without slowing production down.
How a data integration sprint works
- Call (30 min). You show us where the data lives and which question you can't answer today.
- Fixed quote. One page: sources, what lands in BigQuery, what report or app you get, price, dates.
- Two-week sprint. Ingestion on Cloud Functions or Cloud Run, a clean BigQuery model, checks that catch bad rows before they reach a report.
- Handover. Documentation, alerts, access in your own Google Cloud project. You own everything. We can stay on for the next sources.
What we've learned the hard way
Real files are messy
Dates like June 31, Excel serial numbers in date columns, carriage returns inside text fields. We parse defensively and quarantine bad rows instead of failing the whole load.
Partition by the date you query
Partitioning by ingestion time is the default and the most common reason a BigQuery bill grows. We model for the questions you actually ask.
Schemas change
Sources add, drop and rename columns. Our loaders map by header and evolve the schema instead of silently shifting data.
Every tenant stays separate
For multi-client platforms, every query is scoped to the right customer at the data layer, not just in the UI.
More detail in Loading messy CSV exports into BigQuery: what three years of distributor data taught us.
Questions
Which systems can you connect?
Anything that can give us data: REST and SOAP APIs, CSV or Excel files dropped by email or SFTP, databases, ERPs and route accounting systems, POS and fiscal devices, and legacy web portals with no API at all, where we automate the export the way a person would click through it.
Why BigQuery?
It handles hundreds of millions of rows without anyone managing servers, you pay for what you query, and every BI tool and AI model can read from it. For most growing companies it is the cheapest way to get one reliable copy of their data.
We already have a data warehouse. Can you still help?
Yes. We often start by fixing ingestion that already exists: column mapping that breaks when a file changes, dates that load wrong, duplicate rows, slow or expensive queries.
How long does a first integration take?
A first source landed in BigQuery with a working report usually fits in a two-week sprint. You get a fixed quote before we start.
Tell us what you need connected.
A 30-minute call with Nicoleta, our CEO and lead engineer. You leave with a concrete plan and a fixed quote, whether or not you work with us.
