I work with Credible Data, where we’re building on Malloy - a modern open source language for describing data relationships and transformations. In a nutshell, Malloy is an analytical language that runs on SQL databases, and provides the ability to define a semantic data model and query it.
I thought it’d be particularly interesting to see where Malloy can help with OMOP queries, so I wrote up a short example of using Malloy to model OMOP relationships, joins, measures, and so on. The example uses one of the synthetic OMOP datasets.
I’m curious to learn from the community whether this type of reusable semantic model could be useful to accelerate querying OMOP data, prevent errors, and especially now where we may want AI agents to run such queries directly.
Well, as you can see, you don’t get much response to an open question like that. Unless somebody had the same idea and wanted to exchange experiences, what should people say?
You may want to demonstrate how your solution can solve a particular problem people have wht the current setup, which is running SQL queries for ETLing and for analytics. That’s much more effective in my experience.
Thanks @Christian_Reich ,
There are examples in the blog post - are you saying copy those here? or would it be helpful to look at other topics in the forum to see specific use-cases folks are solving for?