A lot of people that sign up for Definite (https://www.definite.app/) have much of their data sitting in S3 / GCP. Most BI tools need you to pipe that data into a SQL database / data warehouse, but that can be expensive and a pain to set up.

We've seen this enough times that we added some features to directly query S3 without moving the data.

Under the hood, our product[0] uses DuckDB as a data warehouse. DuckDB has some amazing features that made it pretty easy on our end.

Here's how it works (https://www.youtube.com/watch?v=yrrCQnfKEig):

1. Add an S3 / GCP key and secret

2. Write a little SQL to access the data you care about (e.g. CSVs, parquet, JSON, JSONL, etc.)

3. Build a model on top of that data (we automatically create a basic model which you can tweak)

4. Share that model with non-technical business users (if they can use a pivot table, they'll be able to use Definite)

5. Ask our AI ("Fi") questions about the data. Fi can build reports and dashboards on top of S3.

0 - https://www.definite.app/