Sigma Computing

Sigma has the right model.

From the user perspective, the columnar formula editing feels familiar enough, and the learning curve is friendly to anyone with spreadsheet experience.

From an architect perspective, the semantic layer is lightweight and easy to adapt to ever-changing schemas. The UI covers most of the functionality you'll want, and for everything else you can fall back on writing your own code, like cron strings or SQL queries. I let Sigma do probably 95% of my querying, leaving the gnarliest 5% for me to write in their SQL runner.

I do a significant amount of lightweight ELT work in Sigma as well. I like being able to write massive views that I know I'm going to have to update often, and just like Snowflake / Sigma materialize them for me. When I use Sigma to query S3 directly in the SQL runner - usually to POC some new data that hasn't made its way out to the business teams yet - this functionality is really convenient because those queries are not efficient. I just need to get new data in front of the business as quickly as possible.

Visual discovery is lacking when compared to mature tools like Tableau. Sigma needs to continue chipping away at the Tableau features around dashboarding and charting.

Sigma is mostly easy to use, and has a very clean and simple presentation. The data is almost always accurate, and has little downtime for updates. My teams have no problems finding the info they need after I showed them where to look!

The dashboard/site could use with some personality and flair. The drab black/white can be somewhat monotonous. Offering classes/tutorials for how to find certain data sets, and analyze your results would be a huge bonus.