AI analytics is only as good as the context you give it. Without a semantic layer - a unified, shared definition of metrics, segments, and business logic - AI (and everyone else) is guessing at what "active user" or "revenue" means at your company. Data Studio is the analyst workbench where that foundation gets built. Define metrics once. Transform raw tables using SQL or Python. See dependencies before changing anything. Publish what's trusted to your Library. Then get reliable answers from AI
As a non-technical person in Metabase, having verified datasets and predefined metrics that are owned by someone who actually knows what they're doing makes it way easier for me to run the reports i need, and be confident in the answers I get.
I haven't asked Metabot yet, but i'm pretty sure she feels the same.
Great to see Metabase still going strong. I used it a couple years ago on a personal project and I liked it a lot.
The dependency graph feature is what sells this for me. Been burned too many times by renaming a column upstream and only finding out days later when a dashboard breaks. Having that visibility built into the same tool where you define metrics feels right — no more duct-taping dbt + Looker + docs together.
Congrats on the launch! This is a big step forward in making data easier to work with and trust across teams
This is my favorite project that Metabase launched, and I use it every day now. It's a set of tools to run your entire data stack inside Metabase: transforms, definitions, lineage, everything.
Here's how I use it every day: - Write SQL (and sometimes Python) transforms and save results straight to the database: like cleaning up messy user signup data and combining it with referral info to make a new table I can query everywhere. - Define metrics once so I don’t have to rethink “what counts as active users” every time: now everyone on the team uses the same definition. - Create clean tables I trust: for example, a revenue table that I know is accurate and ready for dashboards without extra checks. - Trace numbers back when something looks off: like seeing exactly which transform or question a dashboard number came from instead of guessing. - Catch issues early: if a column got renamed and a query breaks, I know immediately which dashboards are affected before anyone asks “why is this number different?”
Everything in one place.
The "what does active user mean" problem is real and expensive. Every company I have worked at has had at least three competing definitions living in different dashboards. The shared semantic layer approach makes sense - it is the same problem that good data teams solve manually, just formalized. How does Data Studio handle it when business definitions legitimately change over time - does it version the metric or just update in place?
Dependencies are my favorite! Being able to follow the data flow and know what the downstream impact to changes is super helpful!
Very exciting launch. Been using Data Studio for a while, and I love how easy it makes it to build (sql/python transforms + library + remote sync) and correct (dependency graphs) a robust, intuitive data environment that other people can actually use and build from 🎉
You know that moment when someone asks "how many active users do we have" and three people give three different numbers?
Yeah, we fixed that ✌️
We're excited to announce that we've launched a new, simple way to clean up your data structure as you grow. Not all companies needed transformations, semantic layers and metadata curation in the past as agents. However, as agent powered analytics become a primary way for people to work with data, a clean data layer matters more and more. Data Studio is how we think you should create it!
As a non-technical person in Metabase, having verified datasets and predefined metrics that are owned by someone who actually knows what they're doing makes it way easier for me to run the reports i need, and be confident in the answers I get.
I haven't asked Metabot yet, but i'm pretty sure she feels the same.