WEBVTT 00:00.000 --> 00:06.000 All right. 00:06.000 --> 00:07.000 All right. 00:07.000 --> 00:10.000 Thank you. 00:10.000 --> 00:12.000 Are we going to sound check, too? 00:12.000 --> 00:14.000 People can hear on the stream. 00:14.000 --> 00:15.000 All right. 00:15.000 --> 00:17.000 Well, I'm going to start with a confession. 00:17.000 --> 00:22.000 So the real reason why Evan and I are here today has to come and enjoy everything that 00:22.000 --> 00:27.000 Belgium has to offer, the good beer, the good cheese, the good chocolate. 00:27.000 --> 00:32.000 But we're also here to preach the gospel of open source, which I think is going to be easy with 00:32.000 --> 00:33.000 this crowd. 00:33.000 --> 00:37.000 It looks like we have a lot of very open source enthusiast here. 00:37.000 --> 00:44.000 But we're going to be up to talking about Apache supersets and let's kick it off. 00:44.000 --> 00:45.000 All right. 00:45.000 --> 00:48.000 So let's start with some interesting shows. 00:48.000 --> 00:49.000 Yeah. 00:49.000 --> 00:50.000 Hi. 00:50.000 --> 00:51.000 I'm Evan and Resecus. 00:51.000 --> 00:53.000 I'm one of the projects. 00:53.000 --> 00:57.000 So I am here for the beer for real. 00:57.000 --> 00:58.000 But yeah. 00:58.000 --> 00:59.000 Yeah. 00:59.000 --> 01:01.000 You've got more beer cred than I do. 01:01.000 --> 01:02.000 And I'm Max. 01:02.000 --> 01:05.000 I'm the original creator of Apache superset. 01:05.000 --> 01:08.000 You might know me to have heard me place is on the internet. 01:08.000 --> 01:10.000 Talk about Apache Airflow. 01:10.000 --> 01:16.000 If you're familiar with that, assume we have a lot of data and enthusiasm in the room. 01:16.000 --> 01:21.000 So I started at Apache superset back in 2015. 01:21.000 --> 01:24.000 It's almost like 10 years ago. 01:24.000 --> 01:31.000 And since then, I started a company coming to you called preset that helps kind of foster 01:31.000 --> 01:33.000 and offer services in and around superset. 01:33.000 --> 01:37.000 But today, we're really focused about talking about superset. 01:37.000 --> 01:38.000 Yeah. 01:38.000 --> 01:41.000 And of course, one of the things is, you know, global switch day. 01:41.000 --> 01:45.000 If we can talk you into one thing and one reason to come halfway across the globe. 01:45.000 --> 01:49.000 It's to switch from proprietary BI tools to open source BI. 01:49.000 --> 01:52.000 How many of you are paying for superset? 01:52.000 --> 01:53.000 Yeah. 01:53.000 --> 01:57.000 How many folks are already using open source BI? 01:57.000 --> 02:02.000 And how many folks are still paying or your orgs are paying for proprietary BI? 02:02.000 --> 02:05.000 Everybody, everybody's got some Tableau licenses. 02:05.000 --> 02:08.000 And some of what we want to be preaching today. 02:08.000 --> 02:15.000 And the core message is that open source in this area, business intelligence, open source, data visualization, 02:15.000 --> 02:21.000 has gotten extremely competitive due to the work of the community over the past decade or so. 02:21.000 --> 02:28.000 And is extremely competitive and pretty much all aspects of what you might want from a business intelligence tool. 02:28.000 --> 02:35.000 And ahead in many ways, you know, there's a lot of advantages you get from open source that you'll never get from an open source vendor. 02:35.000 --> 02:37.000 And we'll talk about some of that today. 02:37.000 --> 02:38.000 Yep. 02:38.000 --> 02:40.000 So really quickly what we're going to cover is all of this. 02:40.000 --> 02:41.000 What is the product? 02:41.000 --> 02:43.000 We're going to look a little bit at how it's built. 02:43.000 --> 02:46.000 We're going to talk about how you can tweak it, which is the fun stuff with open source. 02:46.000 --> 02:48.000 And then, of course, how you can get your hands on it. 02:48.000 --> 02:51.000 And, you know, talk your people into using it. 02:51.000 --> 02:54.000 The start off with exactly what it is. 02:54.000 --> 02:59.000 If what is a patchy superset, so it is, of course, open source we wouldn't be here. 02:59.000 --> 03:02.000 If it was not open source enterprise ready. 03:02.000 --> 03:04.000 Maybe not everyone here and works out enterprises. 03:04.000 --> 03:07.000 There's probably students and college attendees. 03:07.000 --> 03:10.000 But by enterprise, we mean it is very much. 03:10.000 --> 03:16.000 So a patchy superset is an application that you install and that you can use for you and your team, right? 03:16.000 --> 03:21.000 And open source we have, you know, frameworks and pieces of infrastructure. 03:21.000 --> 03:24.000 We have things like databases, small libraries. 03:24.000 --> 03:27.000 But in this case, a patchy superset is an app. 03:27.000 --> 03:35.000 It's a web app that you can install on your network and serve your entire organization company team with very easily. 03:35.000 --> 03:39.000 And then business intelligence, if you're not familiar with the term, even though it is a, you know, 03:39.000 --> 03:43.000 it's kind of a dated aging term that probably started like 20, 25 years ago. 03:43.000 --> 03:49.000 But what it means really is like data consumption, data visualization, dashboarding, right? 03:49.000 --> 03:51.000 So superset will get a little bit deeper into it. 03:51.000 --> 03:58.000 But superset is a tool that enables you and your team to, essentially, like connect a database, 03:58.000 --> 04:04.000 consumer data, rights of the SQL, if you're not a writer, a sample of dashboard, share it with people. 04:04.000 --> 04:08.000 So it's really about data exploration, visualization, creating interactive dashboards. 04:09.000 --> 04:14.000 Now, of course, being open source, we need to talk for a second about the Apache software foundation. 04:14.000 --> 04:21.000 This isn't a patchy project and what that really means is that no organization, like preset or any of the other people that own, 04:21.000 --> 04:23.000 or maintain it are actually the owners. 04:23.000 --> 04:32.000 Apache owns the repo, Apache owns the trademarks, and they provide a governance model that allows us to responsibly maintain the software and maintain the community. 04:32.000 --> 04:37.000 And they provide all kinds of other resources like legal coverage and all sorts of other. 04:37.000 --> 04:43.000 Yeah, in many ways, like it keeps a commercial organization honest around the open source software that is around. 04:43.000 --> 04:49.000 So you'll see, you know, big companies at some point kind of shift their license and things like that with the Apache software foundation. 04:49.000 --> 04:56.000 There's a really clear governance model around things that keeps everybody kind of fair and honest. 04:56.000 --> 05:00.000 And it's all based on the meritocracy as open source should be. 05:00.000 --> 05:03.000 Yeah, it's all about individuals, not companies, which is great. 05:03.000 --> 05:07.000 So we have to play by the rules of engagement to be patchy way. 05:07.000 --> 05:14.000 This is the slide about me, so I'll already introduce myself, but maybe I'll talk a little bit about the genesis of Superset and how it came to be. 05:14.000 --> 05:18.000 So about, yeah, 10 years ago or so maybe like a little bit longer than that. 05:18.000 --> 05:20.000 I was at Airbnb. 05:20.000 --> 05:24.000 And a lot of the tools that we wanted to use for self did not exist on the market. 05:24.000 --> 05:27.000 So we had tablo licenses. 05:27.000 --> 05:33.000 We look at different BI tools, but like none of the tools that existed out there would totally serve our use cases. 05:33.000 --> 05:42.000 And some of the issues is we had picked these big databases at the time, like Apache Druid, Presto, and Tablo did not play very well with these databases. 05:42.000 --> 05:47.000 And we had a set of like very kind of unique use cases. 05:47.000 --> 05:55.000 So we decided to build our own and like often an open source, like you need someone a little bit more delusional and kind of crazy and think like, 05:55.000 --> 06:00.000 why don't we, why can't I build a new state and new thing from scratch? 06:00.000 --> 06:04.000 That's going to what happened in the REST's history. 06:04.000 --> 06:10.000 Just a little bit around what we do, so part of what we do is we, you know, people use all sorts of databases. 06:10.000 --> 06:13.000 Every other data is a new database that's created. 06:13.000 --> 06:20.000 It's a C% plays very nicely with all of the databases that exist in the world pretty much. 06:20.000 --> 06:25.000 That means like whatever database you're using for your project, your company, you know, 06:25.000 --> 06:28.000 superset can most likely speak to it very well. 06:28.000 --> 06:32.000 And just being open source to, it really fosters, you know, 06:32.000 --> 06:34.000 extensibility, connectivity, and aggression. 06:34.000 --> 06:37.000 So we support all databases. 06:37.000 --> 06:38.000 Next slide. 06:38.000 --> 06:48.000 So if you use everywhere, so one thing about superset is we are one of the fairly big open source project by any open source metrics. 06:48.000 --> 06:53.000 A lot of stars on GitHub, like lots of contributors, super high velocity. 06:53.000 --> 06:59.000 Lots of different organizations contribute contributing to it, which is part of like, 06:59.000 --> 07:05.000 you know, when you make a decision to commit to an open source project or software for the long term, 07:05.000 --> 07:07.000 it's good to know that there's a really strong community around it. 07:07.000 --> 07:11.000 And this is very much the case with superset. 07:11.000 --> 07:15.000 So have anyone to talk about some of the components of the bits and pieces? 07:16.000 --> 07:25.000 Yeah, obviously, the point of the tools to build highly interactive dashboards and allow people to quickly get insight into any of the data that you have available to you. 07:25.000 --> 07:30.000 And but working from the bottom up, there's a workflow that typically happens when you're building these dashboards. 07:30.000 --> 07:32.000 You would start obviously with connecting your data. 07:32.000 --> 07:37.000 So like Max said, we connect it pretty much anything that speaks SQL, but even the things don't speak SQL, 07:37.000 --> 07:41.000 you can connect through some sort of intermediary normally, like a presto, you know, 07:41.000 --> 07:44.000 or drill something like that, some other query engine. 07:44.000 --> 07:46.000 So we can connect all sorts of stuff. 07:46.000 --> 07:52.000 You just put into URI here and then we drop you into SQL Lab, which is a fully fledged SQL IDE, 07:52.000 --> 07:57.000 where you can obviously write SQL queries, but you can share those with your co-workers. 07:57.000 --> 08:00.000 You can get a query history in case you were wondering what you changed in your SQL, 08:00.000 --> 08:03.000 between last week and this week or something like that. 08:03.000 --> 08:10.000 But that's where you basically select all of your dimensions and metrics and then you start to build up your data set. 08:11.000 --> 08:14.000 The data sets, one of the key entries is superset. 08:14.000 --> 08:23.000 We have kind of a data set centric philosophy where you really want to use SQL to kind of build up something that's going to power your videos or power your dashboard. 08:23.000 --> 08:27.000 And we do that through either virtual or materialized data sets. 08:27.000 --> 08:34.000 And then when you create those, we have this data set editor, you see, which is a light semantic layer. 08:34.000 --> 08:36.000 It allows you to annotate your data. 08:36.000 --> 08:50.000 You can add in extra calculated columns and metrics and say what's temporal or what's the primary temporal column, things like that, to kind of curate what is in the data set. 08:50.000 --> 08:54.000 Once you have that, we go into this, which is our chart builder. 08:54.000 --> 08:59.000 And we have somewhere hovering around 40 or so different visualization plugins. 08:59.000 --> 09:04.000 We support any kind of visualization library that's built upon JavaScript. 09:04.000 --> 09:11.000 We have many of them integrated and you drag your columns from your data set to populate the visualization controls. 09:11.000 --> 09:20.000 So you select your dimensions, your metrics, you can apply filters and all sorts of other display properties of any given visualization. 09:20.000 --> 09:25.000 And then once you've built up enough of these charts, of course you drop them into a dashboard. 09:25.000 --> 09:31.000 And dashboards are not just a drag and drop collection of charts, but they also provide a lot of other means of interactivity. 09:31.000 --> 09:37.000 Like if you click on a bar in a bar chart and it's a cross filter that will be referenced in that filter bar on the left. 09:37.000 --> 09:41.000 But it also applies the filter to all of the other charts that receive it. 09:41.000 --> 09:47.000 You can create and share filters in that left filter bar to apply. 09:47.000 --> 09:53.000 But who, well, anything you want to any of the tabs and other dashboard components that are available there. 09:53.000 --> 09:57.000 And then once you've got your dashboard and it tells the data story you want to tell. 09:57.000 --> 10:02.000 Often one of the things people ask for is how can I share this on the periodic basis with my team. 10:02.000 --> 10:11.000 So we have this alerts and reports feature that allows you to send dashboards as essentially screenshots to your Slack workspace or to an email address. 10:11.000 --> 10:13.000 And you can do that on a scheduled basis. 10:13.000 --> 10:18.000 Or you can write some SQL query that'll set some thresholds as like an alert. 10:18.000 --> 10:22.000 And it'll then send the screenshot of the data for your tune to review. 10:22.000 --> 10:27.000 Maybe I'll think it's slightly different take to if you go back to the previous slide on on this. 10:27.000 --> 10:33.000 So I know the talk right before is a DBT talk which is a way to sequence build a dagger SQL. 10:33.000 --> 10:38.000 The next talk after this one is an Apache Airflow talk that's also about writing data pipelines. 10:38.000 --> 10:41.000 So it's not necessarily as linear as it may seem here. 10:41.000 --> 10:44.000 But we assume that you have a database. 10:44.000 --> 10:51.000 Some flat-ish data sets on data structure that are pretty much like typically, you know, large data sets with a lot of column. 10:51.000 --> 10:55.000 Normally, I can make checks and dimensions. 10:55.000 --> 11:00.000 Based on these data sets, you can do exploration, the explorer that we just looked at, 11:00.000 --> 11:02.000 assemble some dashboard. 11:02.000 --> 11:06.000 And then the SQL IDE, sometimes it's part of the workflow, sometimes it's kind of on the side. 11:06.000 --> 11:13.000 It's totally possible to use super send get a lot of value from it without necessarily writing any SQL at all. 11:13.000 --> 11:18.000 But we're pretty kind of SQL centric as a tool. 11:18.000 --> 11:21.000 And if you want to talk about the API and SQL. 11:21.000 --> 11:23.000 Yeah, so I'll talk a little bit about this. 11:23.000 --> 11:32.000 So pretty much so because we're an open source solution and really often it's built by engineers for to support entire teams, 11:32.000 --> 11:38.000 but also engineers pretty much anything you can do in super set through the GUI or the UI. 11:38.000 --> 11:43.000 You can use my microphone, that's position one. 11:43.000 --> 11:47.000 Okay, I'll bring it up a little bit. 11:47.000 --> 11:49.000 Is that better? All right. 11:49.000 --> 11:54.000 So yeah, pretty much anything you can do in the GUI and the UI. 11:54.000 --> 11:56.000 You can do through the API. 11:56.000 --> 12:03.000 So you can define and create dashboards and charts, alerts and reports, all this stuff programmatically. 12:03.000 --> 12:05.000 So good API coverage. 12:05.000 --> 12:08.000 There's the component. 12:08.000 --> 12:10.000 So the building blocks of super set. 12:10.000 --> 12:15.000 So what we use to put super set together are accessible through component libraries, 12:16.000 --> 12:25.000 friend and library, so you can also build your own application using bits and pieces of super set. 12:25.000 --> 12:27.000 All right, I'll talk a little bit about what is pre set. 12:27.000 --> 12:38.000 So I think I think that Robert, a friend, Robert from Altunity gave a talk this morning about how to create a commercial entity around your open source product without selling your soul. 12:38.000 --> 12:42.000 So we're very much in that that mindset as well. 12:42.000 --> 12:50.000 In many ways, like, there's some really good incentives and interactions between the commercial entity around the open source project. 12:50.000 --> 13:00.000 In many ways, the fact that pre set is selling a managed service around super set allows us to go and invest and grow the open source project. 13:00.000 --> 13:01.000 She's really positive. 13:01.000 --> 13:08.000 I'm going to go quickly on the commercial stuff, but we essentially offer a managed service that is kind of, you know, 13:08.000 --> 13:13.000 so if you don't want to run super sound in your own and run your own, be on call for it. 13:13.000 --> 13:15.000 You know, you can use a managed service. 13:15.000 --> 13:17.000 If you want to try it, it's going to be easy. 13:17.000 --> 13:20.000 It's going to take super set for a test drive. 13:20.000 --> 13:23.000 Then we also provide multiple work spaces. 13:23.000 --> 13:28.000 So usually when you deploy super set yourself, you have one super set instance typically, 13:28.000 --> 13:32.000 but we allow you to spin up multiple of these for various reasons. 13:32.000 --> 13:35.000 You might want to put them in different regions for compliance purposes, 13:35.000 --> 13:40.000 or you might want to give different instances to different teams or departments to give, you know, separated, 13:40.000 --> 13:47.000 or you might want to do different workflows like dev stage product and kind of promote your business intelligence assets. 13:47.000 --> 13:49.000 We'll see how this thing works a little bit. 13:49.000 --> 13:51.000 Yeah, so I don't want to go to the here. 13:51.000 --> 13:55.000 I know the audience fairly technical, so we'll look a little bit about how works this, 13:55.000 --> 14:01.000 this super abstract, logical model of how this, this work behind the scene is, you know, 14:01.000 --> 14:03.000 you have a database at the bottom. 14:03.000 --> 14:05.000 Super set is going to the center of it. 14:05.000 --> 14:09.000 It's able to, and it's equal against your database, and it allows you to create also, 14:09.000 --> 14:16.000 try it's in dashboard that you can publish and share with your, with friends in your organization. 14:16.000 --> 14:19.000 Now look at a tiny bit more on the internal. 14:19.000 --> 14:21.000 So there's going to be a line on there. 14:21.000 --> 14:22.000 Here's the back end. 14:22.000 --> 14:26.000 So we have a python back end with slacks and SQL off-commies, stuff like salary. 14:26.000 --> 14:29.000 There's a little bit of pandas to do some protein in there. 14:29.000 --> 14:33.000 There's a thing of more than a hundred and fifty python libraries that are 14:33.000 --> 14:35.000 powering this thing behind the scenes. 14:35.000 --> 14:39.000 On the front end is mostly a react front end react app. 14:39.000 --> 14:43.000 That's using all sorts of data visualization libraries and kind of 14:43.000 --> 14:48.000 bringing that into a very cohesive type place. 14:48.000 --> 14:50.000 So that's a high level, kind of. 14:50.000 --> 14:54.000 So we're north of 10,000 then p.m. components on the front end. 14:55.000 --> 15:04.000 We have a Node.js side card that allows other different operations like asynchronous query operations. 15:04.000 --> 15:07.000 There's some web socket stuff that gets built in there. 15:07.000 --> 15:13.000 So probably not too much is going to shift into Node land, but there are a few of them. 15:13.000 --> 15:17.000 Yeah, it's like, it has become such a large web application. 15:17.000 --> 15:18.000 That's multi layered. 15:18.000 --> 15:21.000 So there's a little bit of Node in here in the back end. 15:21.000 --> 15:25.000 But it's not central to initially what we do. 15:25.000 --> 15:27.000 And it's mostly to manage the web socket. 15:27.000 --> 15:38.000 So that, you know, well, so I wish I had a better architecture. 15:38.000 --> 15:40.000 I let me repeat a question. 15:40.000 --> 15:42.000 What is in a superset? 15:42.000 --> 15:44.000 And what is the actual architecture? 15:44.000 --> 15:46.000 What does it take to run this thing? 15:46.000 --> 15:48.000 We don't have a good diagram handy. 15:48.000 --> 15:51.000 I can say it's just going to on the top of my head. 15:51.000 --> 15:56.000 So there is a Python web server that's running a flash application. 15:56.000 --> 16:02.000 As part of it, there is a salary, a synchronous, kind of worker back in. 16:02.000 --> 16:05.000 There is some optional components. 16:05.000 --> 16:07.000 So there's a metadata database. 16:07.000 --> 16:14.000 Multiple caching layers for results sets and larger results set to export CSVs and things like that. 16:14.000 --> 16:19.000 And then I won't get into the front end because it might take all day. 16:19.000 --> 16:21.000 You can deploy it in a number of means. 16:21.000 --> 16:22.000 You can deploy with Kubernetes. 16:22.000 --> 16:23.000 There's a Helm chart. 16:23.000 --> 16:26.000 There's Kubernetes operator that's coming out. 16:26.000 --> 16:29.000 You can use Docker compose or you can do kind of pip install. 16:29.000 --> 16:32.000 Yeah, the easiest way if you want to deploy it today on your laptop, 16:32.000 --> 16:35.000 you would just clone the repo and run Docker compose up. 16:35.000 --> 16:37.000 And then if you want to know what's happening behind the scene, 16:37.000 --> 16:41.000 you open up that Docker compose file and see what's under there. 16:41.000 --> 16:46.000 There would be, I think, out of the box, there would be some redness and postgres back in front 16:46.000 --> 16:49.000 and stuff and, you know, few bells and whistles. 16:49.000 --> 16:51.000 Yeah, and that's how you get all this stuff out of the box. 16:51.000 --> 16:55.000 But a lot of people obviously besides running their internal business intelligence tool, 16:55.000 --> 16:57.000 they want to, you know, tweak it a little bit. 16:57.000 --> 16:59.000 This is the joy of open source, right? 16:59.000 --> 17:01.000 So you want to make it what you want it to be. 17:01.000 --> 17:04.000 And there's a lot of things you can do beyond the normal, 17:04.000 --> 17:08.000 like, let's just run superset and look at some use case data. 17:08.000 --> 17:11.000 You can, you can use embedded. 17:11.000 --> 17:16.000 So we provide a SDK called our embedded SDK that allows you to install a superset dashboard 17:16.000 --> 17:18.000 as a React component. 17:18.000 --> 17:24.000 So you can pass a guest token from your host application and, basically, authenticate. 17:24.000 --> 17:29.000 And it follows all the RBAC rules and everything and shows your data in situ. 17:29.000 --> 17:31.000 And you can really make it look like your brand. 17:31.000 --> 17:34.000 So here it is embedded in somebody else's website. 17:34.000 --> 17:38.000 And you can feed it and all sorts of stuff to me. 17:38.000 --> 17:40.000 Yeah, I'll say a thing too about this. 17:40.000 --> 17:43.000 But like, you know, if you build a web application nowadays, 17:43.000 --> 17:48.000 there's often going to be a requirement to have a portion of your app that serves analytics 17:48.000 --> 17:49.000 to their user of your app. 17:49.000 --> 17:52.000 So if you're building a SaaS application or any web app, 17:52.000 --> 17:56.000 you're probably going to want to have some sort of dashboard in there. 17:56.000 --> 18:01.000 And we see it more and more that people use superset as a way to power this part, 18:01.000 --> 18:04.000 the analytics interactive part of their application. 18:04.000 --> 18:06.000 And then you can do that through embedded. 18:06.000 --> 18:09.000 Of course, it's got to match the look and feel of your app. 18:09.000 --> 18:11.000 So you can talk about theming a little bit. 18:11.000 --> 18:14.000 Yeah, so we provide CSS templating. 18:14.000 --> 18:17.000 And this is surprisingly one of the most popular things people have come to our website 18:17.000 --> 18:21.000 for is more advice on how to do CSS work, which is something we don't provide. 18:21.000 --> 18:26.000 A lot of warranty on because the DOM of any web application is going to change over time. 18:26.000 --> 18:30.000 But there are a lot of very safe things you can do to theme your app however you want. 18:30.000 --> 18:36.000 You can change colors and logos and kind of go crazy with all of the little granular details of how it looks. 18:36.000 --> 18:43.000 And you can do that one off for a single dashboard or you can create a CSS template and reuse it across your instance. 18:43.000 --> 18:50.000 You can also create visualization color themes that are either categorical or, you know, gradient on a spectrum. 18:50.000 --> 18:52.000 So like hop to cold kind of things. 18:52.000 --> 18:57.000 And if you're using open source superset, you can do that in the configuration file. 18:57.000 --> 19:00.000 And create a bunch of those little color palettes and code. 19:00.000 --> 19:06.000 Or if you happen to be using preset we provide a GUI so you can do that, you know, more interactively. 19:06.000 --> 19:11.000 And of course, CSS customization is not themeing. 19:11.000 --> 19:14.000 Themeing is one of the things we're working on right now. 19:14.000 --> 19:16.000 And it's going to be done hopefully. 19:16.000 --> 19:17.000 And if you must. 19:17.000 --> 19:18.000 And we'll see. 19:18.000 --> 19:19.000 Yeah, we'll see. 19:19.000 --> 19:25.000 But it's a lot of work to do because we're built on so many front end libraries, including anti-design. 19:25.000 --> 19:27.000 It's our core component library. 19:27.000 --> 19:30.000 Each part is a superset in creating a new plugin architecture. 19:30.000 --> 19:34.000 So it will be easier to build plugins for superset with all types. 19:34.000 --> 19:43.000 And then of course, if you want to get involved in the community and build more interesting stuff beyond just a plugin. 19:43.000 --> 19:49.000 We have these superset improvement proposals that are kind of the bigger moves we're making in the community, including 19:50.000 --> 19:57.000 including the extensibility and plugin architecture that we're looking at new permissions models, user preference models, all kinds of stuff. 19:57.000 --> 19:59.000 And these are all done via consensus. 19:59.000 --> 20:00.000 That's the Apache way. 20:00.000 --> 20:01.000 We send them to the mailing list. 20:01.000 --> 20:03.000 Everybody has a big discussion. 20:03.000 --> 20:04.000 They get on board. 20:04.000 --> 20:05.000 We vote. 20:05.000 --> 20:08.000 And it's a very open non corporate process. 20:08.000 --> 20:11.000 So if you're curious about what's on the roadmap of superset. 20:11.000 --> 20:12.000 One easy way. 20:12.000 --> 20:15.000 They're going to go look at what's happening now in the community. 20:15.000 --> 20:20.000 As to go look at our get hub issues and filter on the sit label. 20:20.000 --> 20:28.000 Or you're going to see we have an actual sit board with like dozens and dozens of proposals that are in different parts of the development life cycle. 20:28.000 --> 20:29.000 Yep. 20:29.000 --> 20:37.000 And then the other part you can get involved in people do get involved which makes open source business intelligence a really cool place to work is that we have all these different working groups. 20:37.000 --> 20:43.000 So we have a town hall where everybody kind of comes together once a month, but we have a lot of different working groups that are very ephemeral. 20:43.000 --> 20:46.000 So people are working on some geospatial stuff right now. 20:46.000 --> 20:48.000 Some people are working on security. 20:48.000 --> 20:52.000 There's software quality, developer experience, release management. 20:52.000 --> 20:54.000 All kinds of different little working groups. 20:54.000 --> 20:58.000 And we just break into our areas of interest and then come back and put it all together. 20:58.000 --> 21:04.000 That's kind of beauty of having like you know a mature open source project on a mature framework of the Apache delivery foundation. 21:04.000 --> 21:11.000 We have a lot of infrastructure or you know logistics to welcome just about anyone to contribute. 21:11.000 --> 21:20.000 So we see new contributors every month and then you know contributions of all shapes and size and flavors. 21:20.000 --> 21:24.000 I'm going to talk a little bit about like the future of business intelligence being open source. 21:24.000 --> 21:27.000 I think like the real idea there is like the the future of software is open. 21:27.000 --> 21:30.000 Actually the present of software is open source. 21:30.000 --> 21:36.000 Like more and more you know all the cycles and bits and lines of code that execute. 21:36.000 --> 21:40.000 They and they out are more and more open source. 21:40.000 --> 21:52.000 We see that like in their areas where there's a lot of innovation like the LLM world like not only open source is competitive but it's ahead in a lot of cases too. 21:52.000 --> 22:01.000 And I think that explained by that is explained by the fact that open source is just a superior way to write software and business intelligence and elsewhere. 22:01.000 --> 22:04.000 Some things about our community. 22:04.000 --> 22:12.000 So you know the adoption of super since been crazy we like by all metrics viewer to look at you know downloads and stars and contributors. 22:12.000 --> 22:16.000 There's you know we're very well ranked there. 22:16.000 --> 22:22.000 Another thing the next one is about more about the last city so we've had more than a thousand contributors today. 22:22.000 --> 22:26.000 A lot of steady like monthly active contributors as well. 22:26.000 --> 22:30.000 And tons of pull requests so tons of momentum there. 22:30.000 --> 22:35.000 There's no vendor lock in so that keeps us on us at presets. 22:35.000 --> 22:40.000 You know as much as you might want to use a commercial solution to you don't want to run it on your own. 22:40.000 --> 22:47.000 You can go to a solution on preset by the same time if you're unhappy with your vendor you can go straight back to open source. 22:47.000 --> 22:53.000 And we want to make sure that it's easy for people to onboard but also the off board of our management solution. 22:53.000 --> 23:02.000 Yeah there's a CLI tool that allows you to move all of your super set assets between open instances between preset across both very flexible. 23:02.000 --> 23:07.000 And of course you can't have a talk nowadays without talking about AI for a minute. 23:07.000 --> 23:11.000 Like Max said our repo is very active. 23:11.000 --> 23:19.000 We've merged somewhere north of 2000 P.R.s last year and we need all the help we can get not just from contributors but from the bots. 23:19.000 --> 23:29.000 So one of the reasons we think BI and open source in general are going to win on open source grounds is that not only is the code available but the slack is available. 23:29.000 --> 23:36.000 And everything's just wide open bots and AI in general have all of the context that proprietary software does not provide them. 23:36.000 --> 23:40.000 And so they're better able to answer questions and provide help on the repo. 23:40.000 --> 23:45.000 So we've got bots now that are closing a lot of issues on GitHub. 23:45.000 --> 23:49.000 They're able to answer questions very effectively people say thanks bot closed. 23:49.000 --> 23:53.000 And they're also now helping us review P.R.s and merge and all of that. 23:53.000 --> 23:56.000 They automatically tag issues and P.R.s answer questions on slack. 23:56.000 --> 23:57.000 It's all really great. 23:57.000 --> 24:01.000 Yeah we're getting slowly invaded by more and more agents in the repo. 24:01.000 --> 24:04.000 They're mostly helpful and it's kind of scary. 24:04.000 --> 24:10.000 The pace at which is happening but we have yet to receive our first pull request fully written by an AI. 24:10.000 --> 24:12.000 But it's probably coming this year. 24:12.000 --> 24:14.000 I got a hand in the crowd with the question. 24:14.000 --> 24:24.000 Yeah is there or is there. 24:24.000 --> 24:26.000 So let me repeat a question. 24:26.000 --> 24:31.000 So is there some form of like agent or AI integration in the product as it is today? 24:31.000 --> 24:32.000 Is that the right question? 24:32.000 --> 24:34.000 I think there is room for that. 24:34.000 --> 24:37.000 So what we realize trying to build something that we build. 24:38.000 --> 24:45.000 Some text to SQL kind of agent helping with SQL at preset but the design patterns are not super clear yet. 24:45.000 --> 24:49.000 So it's hard to build a solution that will work for the community at large. 24:49.000 --> 24:54.000 People might want to use certain models or certain piece of infrastructure in different ways. 24:54.000 --> 25:00.000 We're working on extensibility and a big extension framework inspired by the VS code ecosystem. 25:00.000 --> 25:06.000 And then we'll make sure that we have all the plugs and sockets for integration and extension. 25:06.000 --> 25:17.000 So that's our angle on that until the patterns around rag and integrating with AI's is our clear which is going to offer an extension framework for people to build those. 25:17.000 --> 25:25.000 And of course they can open source their implementation and other people can decide which plug-ins or extensions they want to bring into it. 25:25.000 --> 25:28.000 We got like three minutes left. How many slides do we have left? 25:28.000 --> 25:30.000 We're good. We're good. We're good. We're good. 25:30.000 --> 25:34.000 Obviously the main thing is you need to get your hands on it if you haven't already. 25:34.000 --> 25:42.000 How do you do that? The main way we would steer people toward is just go to the website and go to the repo and install the darn thing. 25:42.000 --> 25:48.000 So like Max said, Docker compose up. That's really as simple as it is and then you can get it running locally on your machine. 25:48.000 --> 25:53.000 It might be a little more intricate if you wanted to deploy it because there's a lot of things to secure and configure. 25:53.000 --> 26:00.000 They're set up RDS, Redis, RDS. It might be a little complicated to really get a production instance. 26:00.000 --> 26:05.000 But if you'd want to go on your laptop, get a clone, Docker compose up and you can get running. 26:05.000 --> 26:10.000 You can also try on preset. It's kind of the easy way. If you go to preset the IO, try for free. 26:10.000 --> 26:15.000 If you like it, as I mean you have to go at a commercial solution, you can go back to open source. 26:15.000 --> 26:22.000 So you know, there's just like training wheels there and it depends if you want to be on call for this thing or if you want people to write this software to kind of run it for you. 26:22.000 --> 26:27.000 Now it's free for five users forever. It's not a temporary free thing. 26:27.000 --> 26:34.000 Then of course, find us on Slack. Max and I are both on this Slack along with 18,000 other people. 26:34.000 --> 26:39.000 So your good company comes say hello and ask all the questions you want. 26:39.000 --> 26:47.000 Yes, someone will answer your question. It might be an AI agent. It might be someone from the community or if all falls off then it's probably going to end up being one of us. 26:47.000 --> 26:49.000 We keep the bots to a channel. 26:50.000 --> 26:53.000 Why is it called super set? 26:53.000 --> 26:55.000 Well, it's called super set. 26:55.000 --> 26:58.000 Boys at call super set. We have looked for a name for a while. 26:58.000 --> 27:08.000 And the idea was just to, you know, I think the idea was to being kind of a super set everything you want to do around around analytics and BI. 27:08.000 --> 27:16.000 So slowly we're kind of growing and becoming a super set of everything you might want to do in analytics. 27:16.000 --> 27:21.000 And with all apologies to those joining us on the stream. One of the reasons we're here is we'd love to talk to you all about it. 27:21.000 --> 27:29.000 Show you a demo. Get a little hands on. We're going to be at a table for the OSA open source analytics community tomorrow during those hours. 27:29.000 --> 27:35.000 So come on down and we'll show you in person or you can talk to us after the talk. 27:35.000 --> 27:40.000 We'll be hanging out right around here and are we out of time? More questions? 27:40.000 --> 27:44.000 One minute, there's more time for more questions over here. 27:44.000 --> 27:50.000 I saw at the beginning of the presentation. One of the companies out of the organization are using super set to Microsoft. 27:50.000 --> 27:53.000 So what argument is Microsoft has? 27:53.000 --> 27:56.000 I don't know, but there's a really good video online. 27:56.000 --> 28:00.000 That's about an hour long of everything they've built on and on top of super set. 28:00.000 --> 28:04.000 They actually forked the project and added all sorts of extensions and features. 28:04.000 --> 28:06.000 So you can watch it one hour video. 28:06.000 --> 28:08.000 It's the big team that uses it. 28:08.000 --> 28:13.000 Sorry, the question was, you know, why is Microsoft using super set rather than not basically power. 28:13.000 --> 28:18.000 Yeah, so they've actually, they've got a fork of it and they've been extending it with all kinds of extra functionality. 28:18.000 --> 28:21.000 They've got a visual query builder. The project there's called Titan. 28:21.000 --> 28:24.000 They've got a whole team around customizing that. 28:24.000 --> 28:30.000 I think it's about flexibility customization and like the pace at which you can customize open source. 28:30.000 --> 28:33.000 I don't think even being internal to Microsoft. 28:33.000 --> 28:37.000 I don't think they can really customize power BI to their needs. 28:38.000 --> 28:40.000 All right. 28:40.000 --> 28:42.000 I think we're out of time. 28:42.000 --> 28:45.000 We'll take more questions outside if anybody wants to ask them. 28:45.000 --> 28:47.000 Thank you. 28:47.000 --> 28:48.000 Thank you all. 28:48.000 --> 28:49.000 Thank you online.