 Yeah, thank you. Yeah, I'm Evgeny. I'm from Open Knowledge Foundation and I've been working on the frictionless project like for a while and now I'm proud to present you are our new development It's called friction application. Yes, can you? Yes, yes, sorry Friction application and it's a basically IDE for working the stabilizer and other types of data and currently we call it data management for humans. So Sorry It's it tend to be like the recording application but now we are releasing the beta version and It's only one command two commands you need to run to start working this frictions application It's to install you need to Python to have Python and you need to install frictions application and start It's against your data folder like any on your computer so Be like going to release it is as a desktop application soon, but currently it's like this so The interface so it's basically just a typical idea based on really a lot of examples You already know so it's a file manager. It's a Editor additional panels all this stuff. So I just try to be like Familiar for like everyone so you don't need to like learn it and Like one slide ago. We started the application and first you just getting Your files and it's file manager you can like upload new create folder move files, but just typical file manager and Initially like all files are gray. So basically it's not validated or indexed but Then we're starting by clicking on files and it's it just validates your data just by clicking and it by validating it means that Behind frictions we kind of and really worked a lot on it like we have sophisticated validation system detecting tabular errors metadata errors But here it's just like one click and you're getting kind of like status of your data I'll show what can be like errors but first of all just tables it's getting indexed to a sequo it database and you can do just any like Exploring table or data sorting Filtering whatever so it's just like a normal data viewer You can like open the source View to see what's your like initial cc if it if it was a cc file But the errors Like we wish we didn't have like errors in sorry guys In table or data, but they're like a lot of errors. So the power of frictionless technologies is a Finding the errors, so here's an example of a report showing like Here's a table and it's a it was the metadata types was inferred automatically and then application found this like bad cell which is not which is not a Integer, it's a string, but it should be sorry It should be an integer. So We can Enable errors view. So if you just filter All the rules and and show only errored ones So it could be like 10 millions cc table, but you can just see only errors and Here's like you're clicking getting details and it's it's not like ready, but almost ready If you'd like to clean it like manually You can just do it as like in Excel. So just clicking on the cell Fixing the error saving file okay, so the second like power of powerful thing of frictionless is metadata so Frictionless is based on the frictionless standards It's a metadata standards for describing basically like data description language Describing tabular and other types of data in the application it infers metadata like for you for example For normal files, it's only kind of like information about Like format type, etc. For table or once it will be also schema data types and You can edit this and save and I'll show like later. Why why why will you be like editing metadata like here? It's my table, right? So it's it's if you for example publish this table It will not be like just Table with a random name it will have kind of like human readable description. What is it? So Next one. It's a creating charts. So you can click on the chart button Like for example, you were on the table cars and you're clicking it It tries to create like to guess what you're going to like show From your data from data types and it just created this like chart charts like Vigolite so we always try to rely on Existing like standards technologies and So it's it's the first guess and and You can start like editing it currently. It's like really basic Vigolite editor Visual one. So here we kind of like said that we are interested It's not shown here who interested like in brands. It's cars basically and we'd like to see like average price For each brand so it's now this this table makes like more sense than the first one, which was like just random something so But if you an expert The same as with tables you just start Source for you and edit and can edit like your Vigolite by hands So it's just totally open Not yet done, but of course Having all the like data indexed in the Like database of course the application Provides and just just almost ready provides SQL interface so you just query basically so you have you have like a folder on your computer with a lot of ccs You open it and you can like query mere join like your data like just Kind of like secure interface on top of similar to that asset project So it gives you A field to write your community also it suggests Fields and tables so you can click For example on the ID double-click and it will be inserted into the query So you don't need to remember all your data made model And by the way names here Is kind of like the ones you edit as a metadata so you can provide your names It's not from cc so you can make it more like having more sense Also coming but Once you have kind of like a sequel you the same as in the databases You can use it for example for charts if it were just kind of like a cc so you can For the chart it will be not It will be basically the same so you can you can have some files create kind of like if you've used use it in the charts and by the way Also, it will be difference for regarding charts to we go right that here It will be it will support not only we go right supports. I think like cc and Like Jason, but here you can use like Excel and because it's it renders Internally the chart if you work for Excel or like view, which is not supported of course by Vega light so It's a little bit like specific but we have a kind of like big community of people creating standard metadata based on our standards here for example for open data portals and They used to kind of like edit Metadata by hands. So here you can just like Open table schema. It's one of our standards to describe a cc file and just edit it as a Invisual interface and just save it and reuse It's a simple thing of course markdown editing so you can write markdown documents, but more interesting That we're working on the other project of frictionless called Weimar it's a visualization static site generator based on markdown, but with tables charts visualizations and Here you kind of like you'll be able to just you know that I can drop your chart here and it will be creating like Website like this and you can like publish it for example to netlify or github pages and It's just static site. So it doesn't require any like server infrastructure and It's yeah, it's Other files, so usually that assets. It's not only tables and cetera like descriptions PDFs Pictures, so it's kind of like MIT lights and cat Going like to our data set. It's good. Not creative commons. Okay So the same metadata so data and you add metadata to describe your like intention So Finally we creating a data package that about cage is Format to describe a data set so you just create a data package and add What you'd like to include in your data set? So for example like these files So it's here like it also validates. It's valid and we can we can publish it currently currently only second supported but of course like It's because it's based on the friction is framework in python. We already have like connections to other Platforms, so we're going to add it really soon But second this Providing epik a Let's publish and Here we go. It's like on second But it's valid you pretty sure because it showed you like errors and well described Which is also like really important because there are a lot of cc files on all this like open data portals But you don't know what is it and in that the package kind of that the package It is that the package is on it's kind of like API for data so if you have That the package is on somewhere on the web you can read it for example with frictions framework application And in other tools for other languages we have for our for etc. So PM Coming soon This top application other data portals opening like remotes you can for example instance inside the application Faceting classing standard stuff wiki data integration really interesting topic for us and Yeah other Decorative pipelines here. It's really is good to edit manually, but it will be and we have already like a model for it for decorative pipelines So Important notice it's better for next two months. So it's not for production. No for important files I hope you you you all use like gith and like all this stuff, but just just to make sure Releasing in June version one. It's better. It will be version one Frictions application in data literacy can be like really I think really good for teaching. So it's like I have like a kid I'm going to like introduce open refine to him like when he's 16 But I think for frictions application eight, you'll be good, but still say seven years, but Feedback, please don't forget to support Ukrainian people. They struggling and thanks for your time. Yeah Thank you very much for that talk We have several minutes for questions. I see we have a question. I'm gonna give you the microphone Thank you. You have genie for the presentation. It was very interesting It'll be like a Crosco. That was it was a pretty cat in the Russian Sometimes I my question is about we have a friction application. We have frictionless framework and we have the Frictions data specifications as packs and some signs I feel that there's a race between them some some advanced more than others. I wanted to know How do you approach this? First we experiment with the framework and then we update the specifications etc. Cuz I have a very concrete example in the frictionless framework, we have The metadata can be stored in JSON or YAML However, at the specs specification only JSON is allowed, but we find ourselves using more more YAML But that's not part of the specs. So I wanted you to comment some something about How these things go together which goes first? How do you approach evolution of the specs and yes Thank you. I really good one In like really timely because currently we're trying to start working on specs standards version 2 to wrap it up and make it like we're really real industrial standard and Trying to promote it to other open data portals to be kind of like a data API standard one and Regarding all other stuff. It's basically currently it's it's like one. It's one project. So it's the same called a base So they just In sync so when we saw basically frictions application is kind of like just a UI for the for the Python framework So it's it's one project and it's being developed just In sync and that's the one thing Have you seen how it handles? CSVs that get very large like if I Throw I'd saw this on my laptop and I throw a five gigabyte CSV at it Is it gonna hang for 10 minutes or Currently I think it can struggle because frictions validation is not very very fast But we just got and based on your like help also our community member On this idea of indexing files. We got really quick indexer just based on SQLite Like native indexer and if we're going to introduce this like month when you don't validate but just Upload data to SQLite and it's like really fast for local files Yes, it and it's just kind of like view for the sequel the database So it's can be like any size based on if you have like space on your computer Thank you