 Hey everyone, thanks for joining me. Today's talk is gonna be about open source COVID contact tracing apps to improve transparency and adoption. And I don't wanna make you wait anymore, so let's start right now, okay? First of all, let's give a little bit of context. So during these last months, we have seen huge improvements in different areas of to fight COVID, not only related with the development and a lot of open source contact tracing apps, but also related with test methods, best practices to fight COVID, well, to reduce infection, vaccines and so on, right? And it's amazing, all the improvements we are right now seeing at this moment. But today, I'm here to talk about open source and how open source is helping to fight COVID through COVID tracing apps and how this can help to improve transparency. So just a reminder of all these changes that are joining me right now, this talk is not gonna be about code analysis and it's not gonna be about privacy issues or neither a currency test. And just to make sure, I'm gonna focus on software development activity and try to analyze patterns within software development activity in public entities. So the open source code of those apps only. And why are we focusing this analysis in this framework? Well, I'm from Victoria and in Victoria, one of the things we're doing is software development analytics and try to analyze software development activity of open source projects. Just to tell you a little bit of myself, I'm Anna Jimenez and as I said, I work in at Victoria as a software marketing but also I'm finishing my master's degree in data science and also open source advocate. So as you can see, it's like a mix of everything. So I always tend to say that I'm most like an avid with cross-domain superhero powers. So I was saying in Victoria, we are used to do this type of analysis. And in fact, this type of was had some motivation, some inspiration from all the Victorians that were analyzing these apps as well. So we can find Marike. Marike analyzed, I think it was last month. He published the blog post in open source.com analyzing the Corona Warns app development and it's the German, the official German contact tracing app. And on the other hand, Jesus, that it's also another Victorian, published a list of free open source that work for COVID contact tracing with a lot of useful information, not only European content tracing apps that by the way, this talk is mainly is focused on European contact tracing apps. But he published the list and some analysis from all the contact tracing apps such as it was from the apps from Canada, apps from Israel, Estonia, Italy, Portugal, Australia, India and so on. So I will publish these slides later on and you can go and check the different URLs. There's something in common between Jesus, Marike and I. So we are using, we have been using Caldron, this technology to analyze all the different open source projects. So Caldron.io, it's a free open source platform. It's mainly made for analyst, developers and managers. And it helps to understand the community and processes involving software development. And it's also built on top of Grimo Lab. Grimo Lab is the software used by Chaos Project that is part of the Linux Foundation projects. So just in case some of you are wondering how do we get this data from? I want to show you a quick demo of how is it like is to do the same analysis I'm doing right now. I'm gonna perform right now, okay? So this is Caldron. I just went to rather COVID GitHub because I'm gonna analyze GitHub and Git data. And it's just copying this, the URL, create a new project, go to add data sources. Well, Caldron also supports Meetup and GitLab, by the way. I'm just gonna focus on GitHub because the apps are being developed in GitHub. So if you go to GitHub and you got been paced, you click add. And if you wait for a while, you will start seeing all the data. So well, rather COVID for those who doesn't know it's the Spanish, the official Spanish app. And it started, it was launched in September. So it really doesn't have a lot of history. So I'm just gonna go to the time picker and go to September. Okay, so here you can start seeing the data. You can see you have pretty fine metrics. So related with Commits, Incision reviews, reviews are pull requests in terms of activity and then in terms of community, right? So, and also if you want to go deeper or create your own visualizations, you can go to data workspace or more details. Both of them will drop you to elastic search to Kibana. So that's where we are making the customized visualizations. And here you can build your own, okay? So that's how it is, that's the process I have been following and on all the analysis I've made. Okay, so coming back to the slides. I know there are a lot of tracing apps in Europe, not in European Union, just to make sure just Europe area. But I focus mainly on this bunch of tracing apps. So I will be analyzing UK, Italy, Germany, Portugal, Netherlands, Ireland and Spain. I know friends have won. I know, yeah, friends have won and it's not here. Switzerland, it's also have won. That's also have won. But well, I don't want to make this talk never-ending talk, a never-ending talk, so that's why I had to choose. Okay, and I'm really happy to share with you once this talk ends, more details about, sorry, about all the apps. But for now, let's just focus on this one. So these are the names of the tracing apps. And then also it's important to know that all these tracing apps have different times when they were released, when they were actually launched into open source. So as I was saying, rather COVID was released in September. It's a really early, really new open source. But then you can find all the apps that are running since March in many or NHS COVID-19 from UK. Just to let you know that, of course, the analysis will change between apps because I will be having different time pickers when showing the visualizations. So in order to start the analysis, my goal here was to try to identify patterns, try to identify activity patterns. So for that, I make some questions, questions such as, is there a community around these projects? Or how is the evolution of active developers? How responsive is the community? How is the onboarding of those developers? Or when, where do developers work? So let's go step by step. I'm not showing the whole analysis here, I'm just trying to get some of the insights that I found interesting, okay? So the first interesting thing I found out was about the development of some of these open source apps. So for that, I think this visualization is really interesting and really useful because you can see the lines added and the lines will move over a period of time from comments. So even though RIDAR COVID has been posting several comments from September or long September, you can see that the main launch has been, the main release, sorry, has been right just in one day, like the first day, because the rest of the days were almost nothing. And this not only happened for RIDAR COVID in Spain, but it also happened for COVID tracker that it's the iris, the iris official app. So you can see it was only one big pick of lines added, one comment. And that's all. This means, this can tell us that these apps have been working on private repos first and then open source everything using a single comment. And okay, and what does that mean in terms of transparency? You will say, okay, so that means that there is no gate history available. And that means that probably the transparency is not the best one. It's a low level of transparency. Fortunately, this is not a trend. So you can see that other apps had more activity and they were having all the major releases, for instance, Corona one app, Germany, that is from all these list of apps, the one that has more activity overall in gate. You can see that they had been added a lot of lines, not only one pick, but they have three major picks in many Italy, Corona Mildred from Netherlands, Stayaway COVID, also three major picks, and NHS COVID-19, sorry for this, it's 1909, and UK. So let's move now into the onboarding analysis. So this was related with how well did those apps onboard developers and the evolution of developer onboarding over a period of time, right? So for this, we need to analyze gate. And to explain this, I'm gonna show you this visualization, also available in cauldron. So here you can see two sections, the first section, the dark blue, so it's when people made their first commit, so the evolution of those people, the newcomers, and then the lightest blue, it's the people that made their last commit look, so it's when people made their last commit contributors leaving. And this is an example from Radar COVID, this Spanish official app, and you can see that initially, they had a lot of newcomers, of course, but then they are losing it somehow. Nonetheless, I have to say that these apps was open sourced in September, so it's a really early stage app, so let's see how it evolves. The other apps, they're not following the same pattern, so for instance, Corona War app, and from all these layers, the most advanced one in terms of open source activity, and you can see here that they started to have a big level of onboarding, then they reduced it, but now they are starting to gain new commerce again, right? Even in Italy, it seems like when they make the big release, okay, it seems like they had a lot of onboarding, but then it's almost zero, and they're progressively losing new commerce. Netherlands, it seems now they're losing a little bit, but it remains more or less stable, and this is an interesting case, it's Ireland, okay? So here, this is the git commits they made of a period of time, as you can see, there's not a lot of comets, like it seems like they just have been like, okay, I made a bunch of comets here, a bunch of comets in the middle of August, and then now a little bit, but you cannot see an evolution and actually constant git contributions to the app, right? And you can see like, they began to have new commerce, but now they're totally losing them. And more apps, stairway, COVID, Portugal, and NHS COVID from UK, it's an interesting case, because here you can see that at some point, they didn't have new commerce at all, so they completely lose new commerce, and now they're somehow increasing a little bit, so let's see how it goes. So moving now to community response analysis, here I will try to focus on, see the median time duration of reviews, so our public request and issues, and try to see how the COVID's community responds among the different apps. So the first scenario is from COVID tracker in Ireland. As I mentioned earlier, I saw that COVID tracker is mainly based on just doing releases and that's all, not creating a community. And here we can see, it's a clear example, because when I went to analyze the median time to close issues or the median review duration from public request, I just found that they didn't even made any issue at all since June, that is when July, sorry, that is when this app was released as open source. So this is telling us that they are just focused on releasing code, but there is no community involved actually. We can see other scenarios in Italy, I'm just gonna skip this a little bit because we don't have too much time. So here you can see the median time to close from the different tracing apps, the median time to close for issues created, right? And the median review duration of the pull requests. Which is, so let's go a little bit faster. Okay, now, how the working hours is behaving in the community and how is the diversity in terms of organizations. So here we can see two kinds of behaviors, right? So we have, for instance, Corona One app in Muni from Italy and the one of Netherlands that of course during weekdays, you can see that they are doing most of the activity, but they are also working on Saturday and Sunday. So that's in terms of community, this is telling us that they might be companies involved, of course, but there are also individuals working that they might not be from companies, or maybe they're being burned out, that there are people from companies that they have to work also during weekends, who knows? It can tell us as burnouts, like developer burnouts, people that they weren't supposed to be working on Saturday and Sunday and they are doing so, or they can be individuals that are contributing to code during weekends, right? Like open source and suggests and so on. And then on the other hand, we have a different scenario. For instance, in rather COVID or stay with COVID from Portugal, it's clear they're working during weekdays and office hours, and there is no activity, well just a little bit here, but almost no activity during weekends. So this might tells us that is really, really focused. There are only organizations involved, developers from organizations involved in this development. So I also check out the email domains from the different apps. I'm just gonna focus, I'm just gonna show you rather COVID and the next one is gonna be Corona 1 app from Germany because I think you can see the both alternatives, right? So in Spain, the official company that is making the development of this app, even though it's open source, it's Indra is a tech Spanish company. And as you can see, the second email domain that has most activity on gate is Indra. And then inside, it's also from Indra. So it's Indra as well. And you can see that the companies or the email domains that are making the most contributions for this project are Indra itself. But then on the other hand, when you check Corona 1 app development, of course the official organization involved in this development was SAP and you can see it here, right? But you can see that most of the activity of the gate contributions is not SAP. Others is just because there are more email domains. So you can see how diverse is Corona 1 app right now in terms of contributions, right? So you can clearly see the different behaviors, right? In Spain, there are mainly a single organization making commits, making contributions. And Corona 1 app, there are several people contributing from different organizations or from different projects and so on. So just to end up with this talk, some highlights. So the first important thing I would like to remind is that transparency is important in order to build a sense of fair place. First for the contributors that wants to contribute to the project and then it can build a sense of trust among third parties. And then on the other hand, transparency can help to encourage national tech companies to get involved in the development of the app. And with this, I mean that if I've been in charge of developing an app, I'm Indra for instance in Spain, right? I'm Indra and I have to do this privately, then that's all. I mean, all the companies are not gonna get involved. They're just not gonna care about that. But if we open source this, it means that, okay, Indra is gonna be there, but maybe all the companies or the tech companies or the national tech companies can go there, see the code if it's transparent enough and provide some feedback and work together to build a better app and develop a better thing to fight COVID. And that's all. Yes, third story about petroleum. We, as I said, a software development analytics firm and understand how the software that matters to people is being built. And yeah, that's all. Thank you so much for joining me. There you have my Twitter account in case you want to follow me and ask me any questions you want. And see you soon. Bye.