Hacker News Reader

Comment by marcinzm | Hacker News Reader

> Superset is fast, lightweight, intuitive, and loaded with options that make it easy for users of all skill sets to explore and visualize their data, from simple line charts to highly detailed geospatial charts.

I tried Superset a few years back, and maybe it's changed since then, but intuitive is about the last thing I'd use to describe it. Things which I could figure out in a few minutes on any other BI tool literally took me hours of searching. It didn't help that they decided to rename core concepts at some point so half the online documentation made no sense anymore. Others at those companies who tried it at the time said similar things.

View on HN

I also found Superset unintuitive to use and setup as well. I settled on standing up Metabase because it was so simple to get started with trying it since it can be launched as a single jar. The business users loved it and so did I and administration with a Postgres backend instead of the internal h2 database was a breeze.

Metabase is great. It truly is a BI tool. Superset is more of a visualization platform, which works great if you have engineers building reports. Less good if you expect more junior analysts to be super productive.

We ran into the same exact issue with Superset not being intuitive, just for a different audience that is more technical. Also went with Metabase which is good, easy to use, lacks some a few chart types but overall the past year has seen quite a few changes and bug fixes consistently happening.

I just took a look at Metabase.

https://www.metabase.com/demo

Demo is nice.

Yep, we've really liked Metabase for embedding in our platform.

I had the same experience. Featurewise Superset looked better, but after wasting a couple of hours trying to install it, I just gave up.

Instead I installed Metabase in 5 minutes tops: spin ec2 instance, whether and java -jar . I've never looked back.

The only thing that turns me off I'd that it's implemented in an obscure language. At one time I wanted to add some custom postprocessing to an api (given an sql query, get some python/pandas postproc command from a sql comment and execute it in the returned table), but the used language is just not for me (some lisp dialect)

Clojure is not particularly obscure

My experience with Superset was the opposite. It's easy to install using containers. You can have it up and running and connected to ClickHouse in a few minutes. I also found the internal design pretty intuitive--the SQL query lab is much easier than Grafana's editor.

I like Grafana too, but there's basically no isolation between your query and the SQL database at least in the Altinity Grafana plugin for ClickHouse which is the main one I use.

Let’s be honest, intuitive is the last word we’d use to describe most Apache projects.

They are doing pretty well that it is even clear what the project is really even about. Good luck figuring that out within 30 seconds of hitting the average Apache project homepage.

There seem to be dozens of Apache "Big Data" projects that all look kinda the same unless you are a Big Data person.

Even if you are a data person. The ASF doesn't mind overlap between projects [1], it spreads its bets and lets the market choose the winners.

1. https://www.apache.org/foundation/how-it-works/#incubator

Had a similar experience with Superset. A few others have mentioned Metabase and I agree it's better, but if you're looking for a different approach to data, check out Definite (https://www.definite.app/). It's a "data stack in a box". A few things we're doing differently:

1. Built-in data warehouse - We spin up a duckdb database for you to load data to

2. 500+ connectors - You don't need to buy a separate ETL and you can pull in all your data (e.g. Postgres, Stripe, HubSpot, Zendesk, etc.) automatically

3. Semantic layer - Define dimensions, measures, and joins in one place. We have pre-built models for all the sources we support (e.g. the Stripe model already has measures for MRR, churn, etc.)

4. Simple BI - Build a table with the data you want and generate visuals off that table

I'm mike@definite.app if you have any questions.

I've just been playing with superset. I'd have to agree. Things which are easy in SQL are... disturbingly hard or nonobvious in superset.

And the documentation is sparse at best.

It wasn’t fast either when I used it.

What it was though, was riddled with dozens of Python runtime errors and innumerable glitches.

Metabase is where it’s at.

It’s more intuitive than the open source alternatives but is not as intuitive as tableau and others.

Metabase is more intuitive. Also, being unintuitive isn't great but not the worst thing. A project not even realizing that (and thinking the exact opposite) is much much worse. Unintuitive can be fixed with PRs over time. Delusional project leadership cannot.

Are there better alternatives?

Reply on news.ycombinator.com