 Hello, good morning everyone. So I'm Guillem and I'm coming from the Android innovation hub and today I have the pleasure to be talking with Oscar. Hi, I'm Oscar. I I am with Creepix, a data consultancy firm in Barcelona And I'm the lead of the big data cloud and advanced analytics team, all yours. Okay, so the idea today is explaining a little bit how Andorra is integrating this new platform in their routines. So Just to know a little bit who I'm talking to, how many of you know Andorra? It's okay, mostly. I love you. Perfect. So as you know, we are a really small country in between Spain and France and the the inhabitants is around 80,000 inhabitants. So the the key point here is that we receive every year around 8,000,000 visitors. So this is a this is important when you compare to the local inhabitants versus the visitors we receive every year. So when when we start analyzing the economy in Andorra, we realize that the main sector is tourism. So if we we analyze that the percentage of tourism, it represents around 46% of the economy. So when the last government realized this fact, they started analyzing different scenarios in order to switch and diversify the economy in Andorra. So they came up with the idea of becoming Andorra living lab. Why Andorra is a living lab? So mainly because we are really small, so that makes it easy to get connected with people. Then we have access to data. We have flexible legislation as well. And the most important thing is that we are really really close to the stakeholders. So we can set up meetings with ministers really fast. So the projects can really push really fast. So back in the 2015 the government when they realized that they wanted to transform Andorra as a living lab, they created this foundation called the Andorra Innovation Hub, which mainly is is supposed to promote this ecosystem of innovation, then promote open innovation projects, and then trying to reduce this gap in between the public administration, private companies and researchers. And the first thing we made was made an agreement with the MIT. And the MIT is using Andorra as a living lab. So they are testing the projects, they are testing technologies, and all that stuff. But when when when MIT first came into Andorra, different entities in the country, they were not in a collaborative mood. So the big change that MIT made in Andorra was trying to make people understand that they need to share data. They need to collect the data and share it with different entities in the country in order to make this ecosystem of innovation. So after three years collaborating with MIT and we realized that data was really useful for the country. And as we are really small we have this access to data. At some point we realized that we needed a big infrastructure in order to centralize all the data we have in the country and then promote these projects on top of this infrastructure. So here is when we thought that the national data hub was essential for this transformation of the country. So because of no one in Andorra had experience in such technologies, then we went to ClearPigs and Oscar will explain you how they designed this infrastructure. Thank you. So we basically first started by taking the car and driving to Andorra. It's a two and a half hour from Barcelona. And we spent a week with them basically doing discovery sessions talking to Actua, but also talking to Andorra Telecom, FEDA, which is the electrical company in there with the statistics department of the government, and the different companies that would basically become sources of this data hub. And in the end there would also be stakeholders that would leverage it. Okay, so we started talking to all of them. What do you need of this platform? What do you expect? Also with them, basically we collected the requirements. And then we drove back to Barcelona and with the team basically we spent another week first analyzing the alternatives. Okay, so what alternatives do we offer to Actua Tech? We made a couple of phone calls with them. We kind of fine-tuned which ones of the alternatives were making more sense for them. And then basically we designed, we sized this platform and we basically even created a growth plan of how we expected this platform to grow in the future. Then during three weeks there was a certain gap of a couple of months for all the negotiation and internal politics. But then we basically spent three weeks deploying this new platform, installing it, and training and enabling Guillem and his team to be able to leverage this platform to the max for their benefit. And basically since then, that was nearly a year ago, we have been providing our technical experience, our technical support whenever needed for them. And just to give a couple of tips on how we came up with the with the suggestion of the platform, just which are the key points that made us choose what we chose. And the first thing is that it's a multi-organization platform, so it's not only a platform that Actua Tech will use, it will need to connect to different organizations. Each organization with its own security protocols and their own way of operating, which it's challenging. Second, it has to be a platform that is multi-function, so it's not only a data engineering platform. It has to be also a streaming platform. It has to be also a data science platform. It may even be an operational database. It actually is an analytical store as well. The third thing is that in order to boost also the internal economy of Andorra, it was decided that this platform would run on private cloud, actually on a cloud of Andorra telecom, the telco in Andorra. The third one, the fourth one, is that they wanted it to be using open source as much as possible. And the last one, it had to be an extensible platform that they can add new machinery or new more power when required. And that basically made us recommend and deploy cloud data. And for one year, I can successfully say, as you will see now, that they have been very successful in applying it all years now. Thank you, Oscar. So the data hub is, as Oscar said, an advanced analysis platform. And then the point here is that as we are gathering a lot of data from a lot of, for most of the public entities in Andorra, we do have this unique view of the country. And then we are trying to make this exercise of opening the data to the communities and making these transparent action. So another interesting fact of the data hub is that it's acting as a social component. So now we are seeing synergies between entities in Andorra that before they were not existing. So there is an interesting component of an interesting social component in the data hub. So what kind of data we do have? As Oscar mentioned, we work with Andorra Telecom, which is one of our founders. And we do have all the data related to cell phones. So for instance, here in Spain, you have multiple companies, telecom companies, but the phone, Telefonica. So in Andorra, there is just one. So all the data they generate is ingested into our cluster. It happens the same with the electric company. There is just one electric company. So all the smart meters, all the data related to the electric company is as well ingested into the cluster, mobility data. So all the counters around the country, counting the vehicles are ingested into. And then recently, we made an agreement with Visa. So we do have as well data related to transactional data related to the credit cards. And finally, we have as well different sensors deployed all around the country. So we have weather stations, sensors that measure the river flow, and many of them. So this is a unique, if we think in a country scale. I don't think that there is other countries in the world that have such a unique platform and that have all the data unified, centralized and unified in the same infrastructure. So right now, I'm going to start talking about the different projects we are building on top of this infrastructure. So as I said, tourism is really important in the country. So we receive around 8 million visitors every year. The 65 percent of them are one day three visitors and mainly they come because of shopping, shopping, skiing, depending. So when we were talking with the tourism department, how do they understand the behavior of tourists when they are in the country? We realized that till the moment they were using only surveys and methods that probably are not really relevant if we think in the impact of the policies that that can affect the country. So here when we use our data, we can start, I mean the telecom data, when we are trying to mix different data, we can start analyzing special analysis, even temporal analysis. So here we can see different days of a month and by hour, the, I mean it's a hit map, we can see the visits in the country. So all those different analysis will allow us to get a profile of the tourists that visit Andorra. That at the end is what the tourism department use in order to make their business. I mean not their business, but to target the audience they want to get into Andorra. So here for instance, we can see the congestion of people in different parts of the country by nationality. As well, another analysis we can do. So here with machine learning, we were predicting the activity of tourists depending on the zone they were, the nationality, the hour of the day and many variables. So at the end we get a clear image of the profile that is visiting Andorra. So they come in winter because of skiing and shopping and in summer they come mostly because of hiking in the Moutains, etc. Another interesting analysis is when you study the tourists is understanding what do they do before and after visiting a place. So here, for instance, we were analyzing people that were skiing in Gran Balita. Where do they come from and where do they go. So we are adding all these metrics to have a real clear understanding of the tourist profiling in Andorra. So here is the same, but I mean the people who were visiting the ski resort, so where do they come and where do they go. So at the end, all of these, all of these studies are not useful if they are not, if they are not automatized and people cannot use it in a daily basis during their work. So with the data hub, right now we are building pipelines so people can use these all these metrics that we set, but in a daily basis. So for instance, here a little bit what we can see is we get the logs from Andorra Telecom by Splank. Those logs are stored in the HDFS and then through Spark and other libraries such as the Uber H3. We process all this data and we generate our tables that then through Impala we export to our dashboard. So at the end, we can build those tools for the different public entities in Andorra that can be used. In this, in this video, we can see the dashboard related to tourism. So we can see the unique visitants that have been in Andorra from those who are one day tripper, the ones who are tourists, the amount of money they have west, they have, yeah. Then we have a ranking, even if we select a nationality, then we get more information related to one day trippers, the loyalty of those during this period, how many overnight they made, the average of overnight, the total amount of money wasted by related to visa, and even the merchant in which this money has been spent. Another interesting fact that we will see in the video is that the region department was not able to analyze the impact of different events that are organized in the mountains, that are not, I mean that you cannot track the visitors and you cannot quantify the impact. So by using those technologies in this data, we can analyze in this, this example is the ski wall cap that was, that was happening in Soldado, I mean in a mountain. So there was no way to understand the impact. And through the analysis of the telecom data, we can understand the people who is visiting the event. In this case, we saw that we can really see the impact of the event of people from Switzerland, and then we can get even more metrics related to this nationality. So there is also an important fact that, till the moment I've just been talking about temporal analysis, but it's really important as well to include these spatial analysis in the dashboard. So we made a great collaboration with Carto that they are talking next to us with Vodafone. But Carto, they have a lot of knowledge with spatial models and then visualizing spatial data. So we recently made an agreement with them and we are building this tool in order to have all these metrics as I said before, but in a spatial way. So for instance, this video, I'm going to play back the video. So for instance, we already knew as Andorran that people who were coming from France, they just stopped in Pase la Casa, which is a town nearby France. But we never knew the real values. So okay, we know that probably half of the French tourists they stay in Pase la Casa, but this is not a value. So right now, by using those tools, we can quantify this value and we can provide those tools to the stakeholders in order to try to organize new campaigns to attract the French people that just stay in Pase la Casa into the center and visit all the country. So at the end, we are providing tools to stakeholders in order to improve the decision making. Another use case that we are building on top of the national data hub is mobility. It is really important in Andorra because as I said, we receive 8 million visitors every year, which means that there's a lot of vehicles because you can just come into Andorra by car. There is no import and no train. So that means that all the tourists visiting the country need to come by vehicle. So we are approaching into ways the problem. So firstly, we are building a real-time mobility model similar to the Google, but using our own data and our own knowledge in order to improve the information. So giving into the mobility department data in real-time about the congestion happening in Andorra. And then in order to modify the behavior. So the mobility department in Andorra, Tildomona, are using Google. But with this tool, they can really understand what is happening in the country and they can interact with the country. So that means if one street on one road is getting congested, then we can close this road and, I don't know, open other roads in order to solve the problem. The second approach related to mobility is building a mobility model at the country scale. Two, basically create new scenarios and improve the decision making. So how we are facing this problem in terms of building the pipelines inside the data hub. So we are receiving those counters from the different roads. Those ones are ingested by flume, then are get by spark streaming. Then we run a little model and we store the oldest values into Kudo. Once into Kudo, as well as later, it empalages and provides everything into the dashboard that gets needed. So finally, the last use case is related to natural hazards. So in Andorra, there is a lot of snow during winter. So if we add that during spring, we still have a lot of snow in the mountains. And then we have big rains happening. The volume of water that can get into certain moment, it can be really, really high. And during these episodes, there are flash foots, which means that those events need to really be tracked in order to prevent the population. So this is really ugly, but I'm going to try to explain. So we are getting different data from different data sources. The first one is the river flows. So in the main rivers in Andorra, we are getting the flow of water crossing the rivers. Then we have those weather stations that give us, if it's running, and then the snow depth of the different stations in Andorra. And then with image satellite, what we are doing is we are running a model that is probably in the next slide. So what we are doing is trying to understand the to understand the snow coverage of the whole country. So when we add those three data sources, we can understand the volume of water in the country and then make this alert system. This use case is really interesting because before this data drive and approach, when we had episodes of flash foots, we had different entities from the government coming. So one entity was coming with a PDF, another one with a paper, another one with a CSV. So at the end, the idea is we are trying to unify, to centralize and unify all the data in order to provide tools to improve the daily basis of those entities in Andorra. So here as well is similar to the mobility pipeline. We have flume, which is getting all the data from the different data sources. And we have spark streaming that is processing a little bit this data. We are storing all of this in Kudo and then impala is just giving the data into our dashboards. So that's all. It's a pleasure for us being here. I think it's a good approach what we are doing in Andorra at the country scale. I think that a few countries, I mean, I never met someone that is able to say that we have so many data sources in a unified infrastructure. So for us, I mean, it's a great opportunity to promote all these tools, all these data drive and approaches and all these solutions in order to improve the decision making and the daily work by the public administration and the economic sectors as well. So we are going to be outside if anyone has questions. Just come up and we will talk with you. So it's been a pleasure. Thank you so much. Thank you.