[HN Gopher] Apache Superset
       ___________________________________________________________________
        
       Apache Superset
        
       Author : tosh
       Score  : 392 points
       Date   : 2024-02-26 14:36 UTC (8 hours ago)
        
 (HTM) web link (superset.apache.org)
 (TXT) w3m dump (superset.apache.org)
        
       | marcinzm wrote:
       | > 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.
        
         | fayten wrote:
         | 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.
        
           | datadrivenangel wrote:
           | 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.
        
             | sumoboy wrote:
             | 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.
        
           | fuzztester wrote:
           | I just took a look at Metabase.
           | 
           | https://www.metabase.com/demo
           | 
           | Demo is nice.
        
           | c0brac0bra wrote:
           | Yep, we've really liked Metabase for embedding in our
           | platform.
        
         | bushbaba wrote:
         | It's more intuitive than the open source alternatives but is
         | not as intuitive as tableau and others.
        
           | marcinzm wrote:
           | 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.
        
         | codeduck wrote:
         | 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.
        
         | staticautomatic wrote:
         | Let's be honest, intuitive is the last word we'd use to
         | describe most Apache projects.
        
           | pseudosavant wrote:
           | 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.
        
             | slyall wrote:
             | There seem to be dozens of Apache "Big Data" projects that
             | all look kinda the same unless you are a Big Data person.
        
         | atombender wrote:
         | Are there better alternatives?
        
         | FridgeSeal wrote:
         | 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.
        
       | cogman10 wrote:
       | This looks like grafana, right? Why would I use this instead of
       | grafana?
        
         | pachico wrote:
         | The fundamental difference is that Grafana isn't great at cross
         | referencing data in different data sources. (I love Grafana and
         | I pay for the Cloud version.)
        
           | peterleiser wrote:
           | I found that running TrinoDB in a docker container and adding
           | the trino plugin to grafana was very straightforward. TrinoDB
           | feels magical sometimes, except that the SQL syntax they use
           | seemed awkward IIRC. Also, there are inexplicable performance
           | problems with certain queries that require trying subtlety
           | different SQL queries until it snaps out of it.
        
         | prpl wrote:
         | You can't trivially plug grafana in front of any SQL database,
         | and grafana is more about graphing/plotting (usually time
         | series).
        
           | xyzzy_plugh wrote:
           | You can actually plug grafana in front of any SQL database,
           | but I'm not sure it's a good idea.
        
         | samuell wrote:
         | Much more focused on interactive slicing and dicing of data,
         | rather than mostly following a few pre-defined time-series, as
         | is the focus of Grafana.
         | 
         | As such, closer to an open source replacement for PowerBI.
        
         | bfung wrote:
         | grafana is built more for operational and timeseries data, but
         | not so optimal for complex analytical queries. Ex: up-to-second
         | data on cpu load on a host.
         | 
         | superset is the flip side of grafana; not good for up-to-second
         | updates, but good for complex queries. Also, non-time series
         | stuff. Ex: Which customer groups bought which products for all
         | time? <-- that type of BI stuff.
        
         | skadamat wrote:
         | I love Grafana but Grafana doesn't really support non-time-
         | series visualization that well.
        
         | jldugger wrote:
         | They're both washboarding apps, and while I'm sure they each
         | have panel types the other doesn't yet support, I don't think
         | that's intrinsic. The differentiation as I see it, is that
         | Superset is designed to craft SQL queries and visualize the
         | results. The query builder is probably where this shows the
         | most.
         | 
         | To make it more concrete -- coworkers tell me Grafana doesn't
         | work so well with Apache Druid, while Superset supports it
         | quite well.
        
           | jldugger wrote:
           | *dashboarding, yikes
        
       | rongenre wrote:
       | We use this at my ginormous employer in order to give devs
       | limited access to production data.
        
         | rusackas wrote:
         | Maybe you can't say who, but I'm sure curious. Add yourself to
         | this page if you can:
         | https://github.com/apache/superset/blob/master/RESOURCES/INT...
        
       | adlpz wrote:
       | Has anyone tried both this and Metabase? I've used Metabase in a
       | few projects and I find it very nice. This seems more powerful,
       | perhaps?
       | 
       | Is it worth it for BI on small datasets?
        
         | CalRobert wrote:
         | Yes, I am at a company using Metabase, but I have a decent
         | amount of experience with Superset (albeit from many years
         | ago).
         | 
         | The reason we chose Metabase was that it had table joins, while
         | Superset doesn't (unless it has added them since I used it). It
         | also looks a bit sleeker. But I strongly prefer Superset; I
         | found that with Metabase I had to turn a lot of things off to
         | make it usable (Let me see "the_table" not "The Table"!), I was
         | constantly annoyed at the opacity around models vs "questions",
         | etc. and every time I wanted to change a question Metabase
         | insisted on creating a new one instead. The real issue here was
         | when we wanted to swap out the data source for a lot of
         | questions but there was no clean way to do so without MB just
         | creating new questions.
         | 
         | Also, Metabase doesn't have serialization unless you pay them
         | AND you self-host, (if I'm self hosting then what exactly am I
         | paying for?) and that's pretty annoying.
         | https://www.metabase.com/docs/latest/installation-and-
         | operat....
         | 
         | But it does let you join tables. Sometimes that's enough to
         | make MB worth dealing with.
        
           | noduerme wrote:
           | The "model" vs "question" thing is really annoying as there's
           | no real difference from the user's perspective, and it's easy
           | to accidentally convert a model back to a question without
           | noticing when you publish something. You notice when you try
           | to drill into the chart. There's a lot of annoying manual
           | labor in metabase, e.g. I want to filter something into 10
           | different charts and I need to duplicate it 10 times and
           | change a filter on each one. Still yeah joins are nice. A
           | non-bugged aggregate count/sum as a window function would be
           | nicer.
        
           | itsoktocry wrote:
           | > _I was constantly annoyed at the opacity around models vs
           | "questions"_
           | 
           | Yeah, somewhere along the line Metabase decided to get
           | opinionated on "self-serve". I imagine it works well for some
           | teams and companies, but for the tech-oriented, it's
           | annoying.
           | 
           | I prefer my BI tools to be platforms that make for easy
           | charting and cross-filters, while I build and control the
           | models behind the scenes with a tool like dbt.
        
           | adlpz wrote:
           | Thanks! Very detailed answer.
           | 
           | I've found the weird "make it easy" mindset a bit annoying
           | with Metabase too. The whole questions, nice table names...
           | 
           | I'll give Superset a try in my next project I think.
        
           | rusackas wrote:
           | Superset lets you join tables within the same database. If
           | you want to do cross-DB joins, we have a new (beta) in-memory
           | meta-DB that lets you do this, but we generally see and
           | recommend people using things like Trino for this.
        
         | skadamat wrote:
         | Metabase is a bit more user-friendly to be honest than
         | Superset. Superset has a WAY more liberal license, so it's
         | ideal for people who want to customize Superset and build data
         | apps.
        
         | sokols wrote:
         | Metabase is great, I use it with a Oracle Database.
        
       | ldjkfkdsjnv wrote:
       | I remember using Superset in 2017 or so, was forced to by a
       | manager that would not pay for off the shelf software. I also did
       | a few open source contributions to fix some bugs, it was a
       | disaster. A huge rats nest of python. Might have changed in the
       | last few years, am surprised its still active
        
         | jimvin wrote:
         | It's definitely come a long way since 2017! It's improved
         | markedly in terms of functionality and performance. It looks
         | much prettier now as well.
        
       | Maro wrote:
       | I love Superset.
       | 
       | I've been running it in production since 2017, at two jobs, the
       | current one a big corporation.
       | 
       | Best general-purpose, database-backed dashboarding system out
       | there. I would never pay for Tableau or PowerBI.
       | 
       | Same for Airflow.
        
         | atlas_hugged wrote:
         | Same for Airflow? I'm not sure I understand what you mean.
        
           | luccasiau wrote:
           | They were both made by Airbnb and then open-sourced, which is
           | the similarity I assume they meant
        
             | jerrygenser wrote:
             | They were also more specifically authored by the same
             | individual!
        
               | rusackas wrote:
               | Maxime, the original author of Airflow/Superset, is also
               | the CEO of Preset (where I work), so he/we are still
               | working on Superset every day :)
        
       | dang wrote:
       | Related. Others?
       | 
       |  _Open source Business intelligence platform made with Python_ -
       | https://news.ycombinator.com/item?id=29368664 - Nov 2021 (49
       | comments)
       | 
       |  _Apache Superset 1.1_ -
       | https://news.ycombinator.com/item?id=27439939 - June 2021 (28
       | comments)
       | 
       |  _The Apache Software Foundation Announces Apache Superset as a
       | Top-Level Project_ -
       | https://news.ycombinator.com/item?id=25905277 - Jan 2021 (1
       | comment)
       | 
       |  _Apache Superset is an enterprise-ready business intelligence
       | web application_ - https://news.ycombinator.com/item?id=21133931
       | - Oct 2019 (7 comments)
        
       | throwaw12 wrote:
       | anyone knows how does it compare to Looker?
        
         | skadamat wrote:
         | No built-in thick semantic layer, compared to Looker.
         | 
         | I wrote about Superset's semantic layer here:
         | https://preset.io/blog/understanding-superset-semantic-layer...
         | 
         | One popular option is to use dbt or Cube for the semantic layer
         | and pair with Superset: https://preset.io/blog/announcing-
         | presets-ui-integration-wit... and https://preset.io/blog/open-
         | source-looker-cube-superset/
        
       | noduerme wrote:
       | So it's irritating to me that this is ranking #1 on HN (why is
       | it, btw?) I just pulled the trigger on a large data gathering
       | project using Metabase, and feel a bit hampered by the
       | limitations in terms of charts and plugins... but I considered
       | Superset first, and after a lot of thought I decided that almost
       | everything I've ever worked with that was run by the Apache
       | foundation turned out to be semi-abandoned disasterware over
       | time. In fact I wasn't even sure if Superset was still an active
       | project or if it just looked like one, in the way e.g. no one
       | bothered to pull the OpenOffice website offline.
       | 
       | So now that I picked Metabase, Superset is topping HN for no
       | apparent reason. Why?
        
         | hasty_pudding wrote:
         | > Everything I've ever worked with that was run by the Apache
         | foundation turned out to be semi-abandoned disasterware over
         | time.
         | 
         | Amen brother.
        
         | swalsh wrote:
         | Yeah, I'm in a similar thought process. I've been burned
         | multiple times by Apache, will not touch ever again.
        
         | lars_francke wrote:
         | > almost everything I've ever worked with that was run by the
         | Apache foundation turned out to be semi-abandoned disasterware
         | over time
         | 
         | Can you name a few examples?
        
           | bombcar wrote:
           | Here's the list:
           | https://projects.apache.org/projects.html?name
           | 
           | OpenOffice is probably the most famous (it still has the
           | name, but it is dead, LibreOffice is the real "active" fork).
           | 
           | And the things in the "Attic" are officially dead -
           | https://projects.apache.org/committee.html?attic and many
           | more projects should be there.
        
             | lars_francke wrote:
             | I think it's a great feature to have explicit lifecycle for
             | open source projects.
             | 
             | Lots of other projects just die silently and/or you are
             | unsure of the status.
             | 
             | Here you at least have a chance to revive them if you like
             | as there is always an overarching organisation.
        
               | bombcar wrote:
               | The problem really is that some Apache projects are
               | actually alive (Apache itself, apparently Superset,
               | Groovy, etc) and some _appear_ alive at first glance.
               | 
               | More things should move into the Attic, like OpenOffice.
        
           | smaudet wrote:
           | Ivy, Netbeans, Open Office, Shiro, Solr all jump at me off
           | this list:
           | 
           | https://projects.apache.org/projects.html?name
           | 
           | These are all projects that once were (more) relevant,
           | however seem to have become rather niche (Gradle,
           | Jetbrains/VSCode, GoogleDocs/Libreoffice e.g. for the first
           | three are the dominant competitors).
           | 
           | Most of these projects (like the massive commons listings)
           | are either used by some Java library somewhere (meaning their
           | success/relevance is tied to the usage of Java), or are
           | obscure enough that they are no longer used widely and so
           | suffer from lack of interest.
           | 
           | There are gems in this list, to be sure, but if you just run
           | into half-maintained projects all the time you're not likely
           | to associate good things with the Apache name?
        
           | noduerme wrote:
           | Well, OpenOffice as I said. Cordova is/was a hot mess (with
           | some nice pioneering features, just really not well
           | maintained imo and felt like quicksand to build even a small
           | app on) Then the sort of long slow death of Flex (now
           | Royale?) Apache seems like where software no one loves
           | anymore goes to die.
        
             | rpeden wrote:
             | I suppose it depends on projects you're using. For many
             | developers their primary exposure to the Apache Foundation
             | is through projects like Maven and Kafka, and those
             | certainly don't feel dead.
        
         | smallmancontrov wrote:
         | Because we (the FBI Surveillance Van) saw that you picked
         | Metabase, called our shady French-accented overlord, and he
         | told us to dump it.
        
           | noduerme wrote:
           | I knew it!!
        
           | CoastalCoder wrote:
           | I thought your _outrageous_ French accent just meant you 're
           | going to taunt him a second time.
        
           | esafak wrote:
           | He's working on that accent.
           | https://www.youtube.com/watch?v=Z6oeAdemFZw
        
         | smaudet wrote:
         | "semi-abandoned disasterware"
         | 
         | Hmm. I suppose all open source looks that way if it doesn't get
         | regular funding/attention.
         | 
         | Apache does house a lot of abandonware. They had some relevance
         | as recently as 6-7 years ago but they've been largely replaced
         | by nginx I think. That being said, I view them like the local
         | soup-kitchen - important to have and maintain, but not where I
         | want to go for a 5-star meal.
        
           | malfist wrote:
           | The Apache foundation is way larger than just the server
        
             | smaudet wrote:
             | Yes, I agree. However a lot of their forward facing
             | projects seem to be effective abandon-ware (few people
             | interested in contributing, competing more popular
             | solutions based on forks, or just no longer relevant).
             | 
             | These projects don't give the apache foundation an
             | appearance of importance or relevance, rather they make it
             | look rather rundown.
        
               | DaiPlusPlus wrote:
               | That's how open-source abandonware is supposed to work
               | though: the idea is that whenever a (for-profit) company
               | produces something that it can't afford to run anymore
               | but also can't afford to shut-down and damage their
               | customer relationships, then they'll open-source the
               | project and give it to an open-source foundation for
               | stewardship and repo hosting. Yes, it's where software
               | goes-to-die-a-long-death, but it also gives some people
               | hope, and the possibility of giving it a new life in
               | future. Currently, the Apache Foundation is the go-to
               | place for that, and it benefits everyone considering the
               | alternatives are worse.
               | 
               | Obivously the main "alternative" is for the original
               | company to simply shut down the product/service, which
               | can do irreperable harm to a company when they have high-
               | profile customers who are utterly dependent on a service.
               | 
               | Another alternative is to use an open-source foundation
               | that's directly managed by the original company, which is
               | what Microsoft did with its DotNet Foundation (
               | https://dotnetfoundation.org/ ) - and while Microsoft's
               | legal team ensures the foundation is "legally"
               | independent, in practice we know all the significant
               | shots are being called from within Microsoft-proper; but
               | it does give us some modest reassurances that .NET won't
               | suddenly return to being closed-source overnight.
               | 
               | Another alternative is to not open-source it and to
               | instead sell it off to another company that can maintain
               | it while still being profitable - this is what Adobe did
               | with Flash: they sold it all off to Samsung because their
               | Harman division wanted to continue using Flash for
               | embedded/automotive UX work. This approach can work, but
               | doesn't benefit the wider ecosystem the way that open-
               | sourcing does - and something something shareholder value
               | and return-on-investment by selling rather than writing-
               | it-off...
               | 
               | What companies won't do is let any of their engs that are
               | passionate about a project split-off from the company to
               | run and maintain it, le sigh.
        
               | KptMarchewa wrote:
               | I would consider Airflow, Spark and Flink to be their
               | forward facing projects, and they are all very actively
               | developed.
        
           | jakjak123 wrote:
           | Apache hosts many, many projects, some good, some bad, some
           | abandoned, some fucking great.
        
           | nekoashide wrote:
           | Any time I hear "Apache Foundation" my stomach turns as I
           | hesitate to ask my next question. "What we are trying to use
           | from them is built on Java right"
        
             | stuff4ben wrote:
             | That would be anything hosted by the Eclipse Foundation.
             | Either Java-based or abandonware or sometimes both.
        
         | beastman82 wrote:
         | > topping HN for no apparent reason
         | 
         | I think the HN algo is pretty easily manipulated. I worked at a
         | startup that had an effective process to get things to the
         | front page
        
           | CoastalCoder wrote:
           | > I think the HN algo is pretty easily manipulated. I worked
           | at a startup that had an effective process to get things to
           | the front page
           | 
           | That sounds (potentially) sleazy. If you think it's a
           | technique that HN could potentially defend against, I
           | encourage you to explain it to hn@ycombinator.com.
        
             | noduerme wrote:
             | Maybe it's a YC startup.
        
               | ambigious7777 wrote:
               | AFAIK YC startups don't get any more boost on the front
               | page than normal posts.
        
             | ativzzz wrote:
             | > That sounds (potentially) sleazy.
             | 
             | Pretty sure it's as simple as posting in your general slack
             | channel "@here we posted a new article to HN, go upvote and
             | write a comment"
        
         | rickspencer3 wrote:
         | I think that there is an active company behind Superset called
         | Preset.
         | 
         | https://preset.io/
         | 
         | I don't think it's semi-abandoned. I had a brief interaction
         | with the project in my previous job, and I found the community
         | and the company to be reasonably engaged and responsive.
        
         | skadamat wrote:
         | Apache Airflow, Kafka, Spark, ECharts, and many others are
         | still going strong! It really depends on the project to be
         | honest.
        
         | jakjak123 wrote:
         | I have the opposite experience. Lots of good stuff is hosted by
         | Apache Foundation, such as Kafka, Maven, Cassandra, Camel, the
         | Tika project, Superset, Solr, but I will admit they had more
         | relevance 10 years ago. And I dont think there are many
         | organizations that keep open source projects alive longer than
         | the Apache projects.
        
         | tomnipotent wrote:
         | I used Metabase at my last gig (CTO @ e-commerce, 30+ users)
         | and it was well-received and dare I say even a bit adored. It
         | was the only self-hosted tool I'd receive after-hours text
         | messages about going down that someone urgently needed back up
         | for some task due tomorrow.
         | 
         | Business users loved the self-serve query builder, and it
         | wasn't uncommon to walk around the office and see Metabase up
         | on someones screen. My CEO absolutely loved it, and used it
         | daily including to put together data for board decks.
         | 
         | None of my users cared about visualizations, and lived in
         | tabular data. This included finance, marketing, merchandising,
         | operations, and executives (CEO/COO/CFO). The only people that
         | lamented the limited visualization were analysts. Power users
         | did all their day-to-day work in Excel or other tools anyway,
         | such as managing marketing spend or inventory allocations.
         | 
         | Metabase was great for dashboards and self-service (ad-hoc).
         | 10/10 would deploy again.
        
         | renewiltord wrote:
         | Apache Software Foundation is just an umbrella organization to
         | keep things on life support till someone can apply sufficient
         | motive force to resurrect. I think that's really valuable. Lots
         | of projects there have had that effort applied to them and kept
         | going.
        
       | indymike wrote:
       | Can vouch for Superset. I use it in a couple of my companies and
       | love it.
        
       | lars_francke wrote:
       | We've built a Kubernetes Operator for Apache Superset at
       | Stackable: https://github.com/stackabletech/superset-operator/
       | 
       | It's part of our Open Source Data Platform and it's one of the
       | few open source BI tools out there and there are not a lot of
       | alternatives in this space. We generally like it.
        
       | adeptima wrote:
       | Had a very good experience with Superset.
       | 
       | Superset allowed us to replace Tableau and not looking back
       | 
       | Took me a while figure out how to embed it into my app using
       | Superset Embedded SDK.
       | 
       | Superset Embedded SDK - "Embedded SDK allows you to embed
       | dashboards from Superset into your own app, using your app's
       | authentication. Embedding is done by inserting an iframe,
       | containing a Superset page, into the host application."
       | 
       | https://github.com/apache/superset/tree/master/superset-embe...
       | 
       | Superset is based on very high quality and well maintained chart
       | library eChart
       | 
       | https://echarts.apache.org/examples/en/#chart-type-linesG
       | 
       | Community Roadmap
       | 
       | https://github.com/apache/superset/projects?query=is%3Aopen
       | 
       | Huge respect to Preset.io and its team for contributing to the
       | project and keep it in a great shape
       | 
       | https://preset.io/blog/
       | 
       | Superset source code is very easy to read and understand, and as
       | a result it's possible to implement some advanced caching
       | techniques reduce the load on charts.
       | 
       | No BI is perfect.
       | 
       | Watching Superset for years gives me confidence the project will
       | work as supposed down the road, and eventually some of its
       | packages can be reusable for all kind of visualizations and data
       | hacking.
       | 
       | Our main approach to visualisation is to start with eChart and
       | simple Reactjs wrapping and spin off Superset on subdomain for
       | power users, and later see which one works better. Same look
       | gives a very pleasant experience.
        
         | boyka wrote:
         | I have no experience with Superset. Can you elaborate on a few
         | points where you see it excel beyond Tableau?
        
           | adeptima wrote:
           | I dont want to start a rant against Tableau. It's a
           | powerhouse. It's a great superior software. But when it comes
           | to optimizing cost and comparing the total cost of ownership
           | and opportunity to stop paying for Tableau server license we
           | voted in favor of Superset and mix of Reactjs+Echarts
           | widgets.
           | 
           | https://www.tableau.com/products/server
           | 
           | If you have money, dedicated team of data analytics who are
           | already familiar with Tableau - no need to torture them with
           | other tools.
        
             | skadamat wrote:
             | Honestly it's so hard to compare Tableau and Superset.
             | Tableau has every feature and bell / whistle imagine-able.
             | But it's heavy, desktop oriented, and pricey.
             | 
             | Superset is lightweight and open source, but only has 5% of
             | the features. So it really depends what you need!
        
         | Jzush wrote:
         | I'd like to see these types of apps start offering SVG
         | embedding of things like graphs. Frames are such a pain.
        
           | rusackas wrote:
           | That's probably not trivial, but it seems plausible. The
           | beauty of open source is that you can help contribute this if
           | you're fired up about it!
        
         | hughess wrote:
         | We use ECharts in our open source BI tool (Evidence) and it's a
         | great library. Has helped us build a declarative syntax for viz
         | which can be version controlled (https://evidence.dev)
         | 
         | Previous HN discussion:
         | https://news.ycombinator.com/item?id=35645464 (97 comments)
        
           | adeptima wrote:
           | Looks great!
           | 
           | Reminds me Obsidian DataView but with charts
           | https://github.com/blacksmithgu/obsidian-dataview
           | 
           | This whole ideas to have data, visualisations and knowledge
           | base in one private offline place is very appealing
        
         | klaussilveira wrote:
         | How do you deal with data visibility and permissions? I mean,
         | most tables have data that should only be seen by a specific
         | user or group ID, and that layer is usually handled by the
         | application. It would be awesome to expose the power of
         | Superset for users, but I imagine creating the security layer
         | would be a pain.
        
           | re5i5tor wrote:
           | I have this question too
        
             | Ringz wrote:
             | https://superset.apache.org/docs/security/
        
           | spdustin wrote:
           | You can use row-level security, or specify RBAC with pretty
           | much any SQL query.
        
       | rglullis wrote:
       | Anyone that worked it and could compare with Redash?
        
         | skadamat wrote:
         | Well Redash got acquired so development stopped, biggest
         | difference between Superset & Redash. Preset.io supports
         | Superset still
        
           | rglullis wrote:
           | Redash development slowed down for sure, but it's not looking
           | abandoned. It's just that I've been using it for some time
           | now, I'm wondering if is anything feature-wise that could
           | justify the switch.
        
       | vfclists wrote:
       | Generally what you get when VentureCapital/PrivateEquity buys out
       | Redash.io, messes up end users in the process and spits it out a
       | few years later, leaving users confused as to where it stands in
       | the BI tools landscape.
        
       | paddy_m wrote:
       | I wish more projects had guided tour videos that demonstrated the
       | power of the tool in the hands of an expert user. Not "get
       | started" but "why should I care".
       | 
       | Wes McKinney used to have an excellent 5 minute introduction to
       | pandas in this genre.
        
         | skadamat wrote:
         | This might be what you're looking for:
         | https://www.youtube.com/watch?v=kGfUIOK87V8
        
           | paddy_m wrote:
           | I saw that video on the website. It isn't narrated or
           | captioned as to what the users is trying to accomplish
        
         | rusackas wrote:
         | You can check this out. This is a Preset Demo, but shows quite
         | a bit of Superset within Preset (which offers multiple
         | instances of Superset as "Workspaces")
         | https://www.youtube.com/watch?v=V0HwGnC1rU8
        
       | mikpanko wrote:
       | Does anybody know why Superset started trending today? Is there a
       | major release?
        
         | remram wrote:
         | Is there more than this single HN submission?
        
         | rusackas wrote:
         | There is a major release on the horizon (4.0) and there were
         | just a couple of patch releases for the 3.x variants. I'm
         | surprised to see it trending too, but I'm happy about it. More
         | people need to know that Open Source BI is here, and here to
         | win.
        
       | zX41ZdbW wrote:
       | Superset is powerful, but I wonder why they don't fix
       | "papercuts", e.g., misaligned pixels on a spinner, or inability
       | to copy a value from a table's cell, or non-monospace font for
       | numbers in a table, etc. There are hundreds of small annoyances
       | in the product.
        
         | rusackas wrote:
         | We try! We also accept PRs and Issues if there are things
         | bugging people, of course. It's always a balancing act between
         | building some new feature that people are clamoring for, or
         | fixing those cosmetic issues that always crop up.
        
       | posix_monad wrote:
       | Is this capable of performing efficient JOINs across non-
       | homogeneous data-stores?
        
         | Lucasoato wrote:
         | Should it? If you really need that, join the different sources
         | with TrinoDB (or any related managed service like AWS Athena)
         | and connect it to Superset.
        
           | ildjarn wrote:
           | It's common for business questions to only be answerable with
           | a join over a few different stores.
           | 
           | I think Athena can only query data on S3?
        
       | nvrmnd wrote:
       | One thing to keep in mind with BI software is that the users are
       | often very different than, well, those individuals that prefer to
       | use mutt as an email client.
       | 
       | Many, or most, users for a BI tool will be operations, product
       | managers, and business management who simply will not find the
       | interface to be intuitive, responsive, or well designed. At least
       | that's my experience.
        
       | fuzztester wrote:
       | Wow, those Apache guys have so many projects. Of course, they've
       | been at it for years, starting with the Apache web server, then
       | Tomcat, etc., and also, many projects were first developed
       | outside and then handed over to them, for whatever reasons.
        
         | andrewshadura wrote:
         | And sometimes projects are handed to them to die. The way they
         | (mis)handle OpenOffice is unforgivable.
        
       | emilienaples wrote:
       | How would you compare Superset with PowerBI for analytics and CSS
       | integration? Trying to develop features and advanced analytics
       | capabilities into an app?
        
         | rusackas wrote:
         | You can style dashboards with CSS as much as you'd like, though
         | there are some limitations (canvas/webGL elements). I wrote a
         | whole blog post on it: https://preset.io/blog/customizing-
         | superset-dashboards-with-...
         | 
         | If you want to style the whole application, you can fork the
         | repo and go bananas. If you're looking for theming, there's
         | more to be done yet on that front, and I wrote an article on
         | that too: https://preset.io/blog/theming-superset-progress-
         | update/
        
       | Wilduck wrote:
       | Is Superset a decent tool if you're just a single person doing
       | data analysis? Say I have a handful of sqlite databases, and just
       | want to be able to develop some queries / charts. I was looking
       | into Tableau / Power BI / Superset, and all of them seemed pretty
       | heavyweight for a single user, and none of them seemed super easy
       | to get setup locally.
       | 
       | Any recommendations for a good piece of software for the single
       | user case? Or a more convenient way to run the heavyweight tools?
        
         | unixhero wrote:
         | Tableau is the best, most powerful, most mature of the three,
         | most feature complete and easiest of the three. I think they
         | give you a 30 day trial.
         | 
         | This is a single user application, unless you make it part of
         | your built application.
        
       | spdustin wrote:
       | For my last employer, I set up Superset for a number of our
       | clients to show all sorts of heavily customized marketing
       | analytics dashboards, web performance graphs, project management
       | burndown reports, you name it. As with another commenter's
       | experience, we also got a client to replace Tableau with it, and
       | not look back. Such a great product.
        
       | twic wrote:
       | How does this compare to Jupyter notebooks and the ecosystem
       | around that? Do the use cases overlap, or are they completely
       | different things?
        
         | Lucasoato wrote:
         | In my experience, people with a business related background
         | have an easier time learning how to use BI tools (this is true
         | even if Superset may be less user-friendly than other
         | commercial product like Tableau); Jupyter is an interactive
         | computing platform that is based on notebooks and cells, that's
         | more useful for data scientists/engineers whose needs might
         | exceed the capabilities of a SQL interface.
        
       | tomrod wrote:
       | It's been a few years since I evaluated superset. Did they ever
       | resolve drilldown (filter for one chart on a page, populate to
       | all charts)?
        
         | rusackas wrote:
         | Yep... there's Drill By, which is more flexible than drill-
         | down. Rather than having to specify a strict hierarchy of
         | drilling "levels" you can pick columns, hierarchical or
         | otherwise, to drill into.
        
       | rietta wrote:
       | Neat. I have to admit I about had a heart attack reading
       | "Superset" as "Sunset" at first. I've become too jaded about
       | stuff being shut down and announced on HN. Very pleasantly
       | surprised when I read correctly and clicked through to see its
       | about data analytics.
        
       | cheema33 wrote:
       | I recently discovered Apache Superset. I would love to use it in
       | our product. Does anybody know if it possible to integrate it
       | into an existing product? I am mostly curious about hooking up
       | its authentication system to our own authentication system, which
       | is based on auth in ASP.NET Core 8.
        
         | Cilvic wrote:
         | >Took me a while figure out how to embed it into my app using
         | Superset Embedded SDK.
         | 
         | Superset Embedded SDK - "Embedded SDK allows you to embed
         | dashboards from Superset into your own app, using your app's
         | authentication. Embedding is done by inserting an iframe,
         | containing a Superset page, into the host application."
        
       | HermitX wrote:
       | Here is a fantastic video made by Soumil Shah, using
       | MinIO+Hudi+StarRocks+Superset. It is amazing to have an
       | interactive query experience on a data lake directly!
       | https://www.youtube.com/watch?v=JkKBzrQTKx0
        
       | atbpaca wrote:
       | This looks really good! How does it compare to Tableau?
        
         | rusackas wrote:
         | Well, it's free! Or significantly cheaper even if you opt to
         | use Preset to run a hosted/managed/compliant version of it, and
         | not have to deal with config/security/upgrades/migrations. This
         | article is a year old, but it might help a bit:
         | https://preset.io/blog/apache-superset-vs-tableau/
        
       ___________________________________________________________________
       (page generated 2024-02-26 23:00 UTC)