 Cool. Thanks so much for everybody joining. I'd love if you could all just go around and introduce a little like introduce yourselves and tell our audience a little bit about what you do at Get Love. I'll start. My name is Emily. I do mostly content and social marketing on our growth team. I'm kind of an interloper here among our data team on the call. I'll jump in next. I'm Emily. I'm the data analyst on the data and analytics team. And I'm Thomas. I'm the data engineer on the data analytics team. I'm Taylor. I'm the manager of the data and analytics team. And I'm Jacob. I'm a staff engineer on the Meltano team slash lead on the team. And Josh Lambert product manager for the Meltano team. And Chase Wright, FBNA lead on the finance team. Cool. Awesome to have all these different perspectives here. So this might be a question for Jacob or and or Joshua. I just love if you could share a little bit about what Meltano is. I'm at sort of a high level and we'll get more into the details later. Yeah sure. So Meltano is aimed to be a complete solution for data scientists, data analytics people, anybody who's using data. And so you know the complete Meltano actually stands for something where you have Meltano model extract load transform analyze notebook and orchestrate. So you're taking the complete life cycle of the data and you're doing it all in one tool. So where you're modeling it with dbt you're extracting it with our own custom extractors and you're doing this all from one step to the next to to get the whole picture and to make it easy and all kind of in one package. That makes sense. Awesome. So maybe if viewers are familiar with GitLab I read in your read me that Meltano is like the full data science life cycle and GitLab is the full with DevOps life cycles. That's kind of an analogous way. Exactly. Yeah. And that's the parallel and rather than currently the plan is not to wrap it directly into GitLab but to make it a similar thing to GitLab where it is the complete package here and GitLab is a complete package for software engineers. Very cool. And so I think the relationship between like all the people on this call is I think if I understand correctly the data team are sort of like the main customers right now for the Meltano team but eventually you want Meltano to be a fit for like marketers and people ops. I think I saw an example in your functional group update about people opt using Meltano to like see how long it takes to get from a resume to hiring someone. So could you like and I think you eventually want to be able to replace Salesforce and Marketo for those types of for marketing users. Is that is that accurate. Can you talk more about the future kind of scope. Yeah. So scope wise you know this is also very open ended because we really just started to kind of figure out where we're headed. But right now if you take just from a very basic standpoint if you take a data source like lever which is something that our recruiters have and they want to figure out the answer to a problem like how long does it take to get from resume to hire. First you need to extract that data out of lever. So you can extract that data out of lever with our extractors and then you need to load it into a data warehouse. You need to transform it with dbt and you can do all this potentially through a user interface. You can potentially automate all this. And then once you write the melt files which will allow you to visualize all that data you're going to be able to actually see the answer to your question in charts and in tables and actual answers to your questions where you know before it was like SQL queries and raw data that gets transformed into actual visual things that could even be dashboards that go on some sort of leaders screen in their office. Awesome. Thank you so much for sharing. So now I guess enough of putting Jacob on spot I'd love to just open it up to everyone if you want to talk a little bit about or the data team if you'd like to talk a little bit about how you're using it right now and how maybe that's informing the development of it going forward or anything else you want to share. Yeah for sure. I'll have been here. So as a data and analytics team our kind of charge is to set up a data warehouse, pull in all of the data from all these external sources and then present it in a usable manner for business users to gain insights from all that data. And then long term it's to be able to make predictions based off that data and saying okay there's a trend in our sales or marketing or whatever. Why is that happening? And we can start running experiments and getting that feedback loop. And that's kind of where the data science aspect comes in and saying okay we're going to make a prediction by changing this lever and we have the data to track it and measure it and to understand the effects of the changes that we're making. And so our relationship to Meltano is basically saying you know we're going from zero to one here and so kind of when I came in there were some basic there was like some of the groundwork was already laid but there was a long way to go and still is and so it's it's been about getting data from external sources so Salesforce, Marketo, Libre all these different extractors together moving it into our data warehouse and then we use a transformation step and we use a tool called dbt to kind of model the data. And then right now we're using a tool called Looker to visualize it but in an ideal world all of that would be done you know like open-source tooling everything would be version controlled you'd be able to understand and track the state of your entire analytics pipeline from raw data to visualization completely you know within a get tree history essentially. And so that's that's what got me excited about coming to GitLab is basically saying well GitLab is awesome at DevOps data people are terrible at DevOps practices and we're like 15 years behind where software engineering is I think and that's that's kind of why I was super excited to join GitLab was to kind of bring the best of DevOps to data and analytics. That's awesome and you think that's like a it sounds like that's a really common thing like your peers who are data scientists or data ops people would they hear that and be like oh like yes it's like really nice now that there's a solution for that. I think so yeah I mean there's we're attacking a lot of the kind of business operations side of the data and analytics data science world. There's a lot of other data sources larger scale different different kinds of questions but I think we're trying to come up with the the right sort of primitives for the field in terms of how to think about the data problems how to structure the tooling so that we can get some collaboration from other people saying hey like we're going in this direction how are you using this what are your data problems this is you know these are the problems that GitLab has but we want to make this tool useful for a lot of people. And we're just taking whatever they say because you know they are the data experts and you know we are the software development experts and so we're trying to merge those two ideas and just bring the two you know the best practices of the software development world into the data analytics world and so like lots of really great feedback from them to say like this is what we really need and this is what the tool should have and this is the current problems that we have and this is how we can solve it and it's great for them because we build whatever they want. That's so cool. Are you on like have you iterated several versions already based on their feedback or can you fill us in on like kind of how the development's going and what that feedback loop looks like? Can you give any examples of things that they asked for? I mean so I think in the beginning it's just catching up it's like we have a lot of we have a lot of data sources that we needed to create extractors for and we were just like heads down creating a lot of different extractors so we've created ones for lever and net suite and Salesforce and each one of those is like a major undertaking and the other thing is it's also a blank canvas and so now we've identified like this is how we're going to make extractors and we're gonna separate the extract part from the loading part and we have like a clear definition but you know right from the beginning we didn't always have that clear definition and now we have like a really a really solid way of building these extractors and so while we were heads down for a while now we have I think six or seven different extractors that we can pull from and now we're putting that all together into like one kind of MVP in this release. Yeah I think we've been sort of learning as we've been going here so for example Jacob mentioned the split of concerns of extracting data to loading data into a data warehouse you know for example our current data warehouses we found wasn't going to scale to some of the needs of the company and so we were like oh we should probably split these two things out so it's easier to add additional sources and the same thing goes as we go along here and learn for example about some data sources that have very low limits on how much you can pull from them and so you need to sort of have a method to pull small bits of our time and have a way to like configure like backfill and things like that and so we're kind of learning about some of these needs and essentially as we've been designing you know the first extractor was very basic and as we've been going we've been sort of learning new features and new requirements and better ways to do these and so we've been getting better and better and better with each one as we've been going. I think I think the feature that I'm excited about and the one I tell people about the most kind of is the review app feature on different branches so currently what we're doing is whenever you push a new branch to the project you're actually it'll create a clone of the data warehouse and so you'll have your own instance of the data warehouse to go in. You can add tables, delete schemas, mess up the data, do whatever you want and once kind of the pipeline passes you can merge that into the production branch and then your job will run on production and it's kind of a you know something that's pretty standard I think in the software development world but the idea that you would kind of have your own safe data set to play with is relatively new and I think the other big feature is we use GitLab CI as the orchestration tool so state of the art right now is a tool called Airflow for running your data pipelines, moving data, transforming it and running different summary statistics on it but we're actually using GitLab CI and kind of pushing it to its limits to see you know where it needs some work for kind of data people but there's other ways where it's I think better probably than Airflow in terms of frequency and just being able to trigger it based on you know different code changes so those are the things that I'm I get really excited about and and the cool thing is is that now that we're like battle testing GitLab CI and figuring out how we want to integrate it directly into our tool then we say like okay GitLab CI doesn't have this feature or it needs this feature or maybe it has all the features but the UI just doesn't explain it as well as we'd like then we can you know actually make contributions to GitLab CI, did we lose Jacob? No I'm here. Awesome thanks sorry I think I think we caught the end of your thought. Sorry about that yeah the idea is that we've made contributions to GitLab CI and and then you know those would go upstream and this would actually be a contribution to GitLab itself you know to make GitLab the product better as well. Very cool and that actually reminds me of something else that's interesting I think I got ahead of myself but that the I think if I'm correct like this idea rose because we were trying to solve a problem that GitLab had internally which is that it's a lot harder to understand marketing and sales data and so I wondered if anyone could who was more involved in that could explain just how that came about because I just think it's interesting a lot of organizations will have sympathy. Yeah I can I can take a stab at that so with regards to a different perspective like I I'm focused more on like trying to figure out business decisions so you know a simple example would be like you know are we losing money are we making money that's pretty straightforward but it's a really it's it's surprisingly difficult to be able to answer those questions efficiently and when you start getting into more detailed level it becomes even more difficult to answer those questions so what I what I'm really excited about from the Meltana team is to be able to have a solution then I can make a business decision really quickly so to kind of simplify this if my house is too warm and I need to to figure out how to cool it all I have to do is go to thermostat and turn it down that's it a business decisions should be as simple as that and unfortunately it's not so that's what really gets me excited about what the team is doing. I think one of the things I'm super excited about is just making these tools available to everyone and because like right now you know what first of all you don't have this complete picture and I think we're going to make it a lot easier to digest especially for even people who are new to data analytics it's going to be like well this is this one thing that does all the things you need and then also you know being able to like let everybody use it at once versus you know only being able to really have like 30 people on it because for whatever reasons maybe it's too expensive or something like that and yeah. I think another really important thing for me is hopefully try and make these sorts of analytics more accessible to people like you know in a past life I've heard from others is that trying to hook up just some examples of like Marketo to Salesforce to understand you know how a marketing campaign or a conference you attended or something like that you know you invest an X amount of money was it worthwhile is a pretty common question to ask and to try and wire this all up yourself and to get things talking to each other it is is non-trivial and I think anything we can do to make that easier for companies and for users to try and make better use of their data it is a great one so hopefully we can accomplish that with with Peltano here and really kind of reduce the initial friction of getting some of this data and some of the insights out of your own data you're getting from these tools there are whole swaths of small and medium-sized businesses that don't have access to data analytics because they don't have engineers on their team they don't have the reports that they get are the reports they're getting from whatever tool they're using whether that's Marketo or Salesforce or Shopify or whatever the back-end tool they're using is and so the problem is that when they're dependent on these like silo data sources they can't do anything cross-functionally so if you run a giveaway and you get all these new email sign-ups and they're piping into MailChimp and you want to know if those users are buying things in Shopify if there's not some sort of native integration you don't know and you can't relate that data to any other data source I think Peltano especially as an open source project can make it really easy for all these companies that don't have a ton of money to invest in data analytics which is a new field to a lot of organizations they will finally have access to the data through you know the cost difference that Peltano can be to them that's so awesome to hear and so it was great we talked a little bit earlier about how the vision is that this will be an open source tool as well so I'm just curious on behalf of our readers and viewers are you able to accept contributions already and are there any like types of contributions that you're most you know in need of or prioritizing right now we are absolutely open source right now and you can read all the issues and you can look at all the development that's currently happening we just put out our plan for an MVC so if you're if you're open willing to help us get to that MVC then that is that's the next step for us but currently we have a plan to to get from zero to an MVP an MVP and and all contributions are open and we have many places that you can contribute because we have Meltano analysis which is the UI that's on top of everything we have the extractors which are individual extractors that we write in a certain way you can look at the Fastly extractors an example and then we have the loader is currently we're only supporting Postgres but we are going to need to support many different database types that especially data analytics people use like Redshift and BigQuery and stuff like that so if you want to write a loader for those that would be amazing we're going to do it eventually but if you want to do it before that fantastic and and many other things as well awesome and so we might wrap up a few minutes early but I wanted to give you all the opportunity is there just anything else you'd want kind of your peers data scientists other like types of users like anyone else what it is there anything else you want them to know about Meltano and the work that you're doing I would just also say that the development of Meltano is all done in Python the front end is done with Flask and in the JavaScript is written with view with view X so if you're a view aficionado like myself and you love to write view like I do then it's a really really cool project to contribute to because you're doing a lot of charts and a lot of really cool data visualization stuff that you don't get to do on your basic credit application I would put in the plug kind of for the analytics team so our primary analytics project is in the Meltano namespace it's right now it's Meltano analytics and we're you know we use Meltano but we're also kind of off scrolled away doing our own things to me kind of the business objectives but since everything's kind of out in the open all of our database transformations you know a lot of the stuff that we're doing is right there out in the open what I would ask of the community is like help us think about these things you know there's we know that some aspects of how we do analytics how we do data science is not where it should be it doesn't it doesn't feel right it doesn't have the same level of process and stability around like software development does like with git and murder quest and things like that help us think about this stuff you know I think blog posts will be great to kind of get some of that feedback but you know if you don't think we're using the right primitives or we're going about it the wrong way like we're all ears we want to do we want to be world-class at this both on the Meltano the product and also like the internal data analytics team we want to be world-class of what we do so we're all ears and same team as well if if you look at the tool and you say it would be really great if it did this feature requests start sending feature requests we'd love to hear them where Taylor and Jacob where do you want that in issues should people submit those sort of requests to write in issues should they everything's an issue yeah and I'm gonna make a there's the Meltano group and the Meltano project which is a giant Mono repo right now which is the big plan for the future right now I'm gonna go make a label called feature requests if you want to make a feature request and tag it awesome if there's nothing else then thank you all so much again for joining I'm gonna drop everyone's links to everyone's bios and social and everything at the bottom of the blog post people can get in touch that way too but otherwise this has been an awesome first session I'm glad we all got to introduce the Meltano team a little bit to our to our readers so I hope everyone has a great rest of your day and thanks for joining thanks Emily