Why Use It?
Not sure how to get started with Trilogy?
Checkout the blog series on Trilogy in Practice
The Overall Value Prop
Trilogy - in the long run - aims to build a virtuous ecosystem across 3 core personas - analysts, engineers, and data consumers. The core Trilogy engine is the glue that holds these together, providing a consistent language for information exchange and iteration.
Each user can engage with the ecosystem through a different entry point:
For Analysts
Trilogy Studio and VsCode can be the solution, providing interactive workbenches for iteration. You can import models defined by your central analytics team and write "SQL" against them. When your engineering partners updates their tables, your queries don't have to change, since the model can abstract away the migration. (or make it very easy to compare v1 when v2). When they don't get to your feature request, you can extend the model yourself and then pass that info back up to be moved into the central schemas.
Outside a business context, you can leverage an open-source model store like the Trilogy public models and so you'd be able to get directly to querying models for common use cases.
For Engineers
For data engineers, the TrilogyT package is a suggested entry point - write your models + outputs in Trilogy, compile that to the theoretically most efficient DBT graph, run in your tool of source. Bonus points that your published model becomes the downstream consumption layer. You get a nicer DevX, more testability, and the open-source toolchain.
For Consumers
For general business consumers, Trilogy provides an accessible query language. When even that is too much friction, Trilogy-NLP helps you minimize the quality problems inherent in AI to SQL and query your high-quality semantic layer with natural language.
Visually
TIP
Though these are the suggested paths; the open and extensible Python API means that you can mix and match as needed. Trilogy can be fully embedded in other tools as part of a larger toolchain.
For those who know SQL, Trilogy is a safe entrypoint into a data model. And for the rest, engineers build reporting/models off trilogy models, including natural language tooling to make the data more accessible than ever.
The Value of a Common Language
The value proposition of SQL - that Trilogy strives to preserve - is that it offers a lingua de franca for language across these three groups. Solving for any one in isolation does not provide the exponential benefits of solving for all three.
That's why Trilogy aims to provide first-class tooling across the data lifecycle - analysts should be able to start small and get value as quickly as possible; engineers should be able to take prototypes and specs from the business and production-ize them or design new data applications from scratch with strong quality checks, guarantees, and SLDC; and consumers should be able to get the data they need in the format they need it, without having to know the details of the underlying data model.
How do we know it works?
Ultimately, users get value if it helps them work with their data faster and better. Trilogy is built by people that loved SQL but got frustrated with it; our north star is bringing back the joy of SQL.