 So welcome everybody again to our opening keynote by Felix DeLachra about Open Geodata for Humanitarian Action. Free? Yes, now I'm on the mic. Yeah, thanks. I'm very glad to be here. It's my last, my first talk or like participation in a Debian event in seven years, I think. Wow, time is passing by so fast. In the meantime, I have been doing work with maps, with digital maps and different parts of the world and how to use them for good, for international cooperation and for humanitarian action. My background, I'm, I study digital media. I'm an open source since I'm, I don't know, to 12, 11, something like that. And the last seven years I focused on maps and especially on places where there are no maps. Those are still surprisingly a lot of places in the world and that has a lot of implications to the people that live in those areas, right? If we are in Europe, we are used to have a map and we sometimes even think about why do I need anything else, right? Because I can, I have my, my tools that I use, I have my maps that are printed that I can read. But that's not the case everywhere. And if you're not living in a place where there is no map, you cannot figure out easily how to come from A to B. And it will be difficult maybe to reach a good doctor to know where's the next hospital. But it has also implications on access to jobs, on leisure, right? If you know where to go, you probably would combine your ways more efficiently and you could, you could have more fun in life. And of course, when there's a disaster, then everything becomes like really essential and crucial and important. And in the case of a disaster, imagine that you don't have a way of how to make even a plan of evacuation or where to find the next water supply. But what if one could change that? And of course there's open source, there are all these tools so we can go online, we can use this and add information to it. The parts that matter to each of us, that could be the neighborhood, but it can be also other things in your city or your country. One of the wonderful things that you can do with maps is you can travel and at least in your imagination and that's maybe also better for the carbon footprint. So I would like to take you to travel with me to some places in the world and to see what people have been doing with maps. And I want to start with you in Managua and Nicaragua where I've been living for a long time. This is not directly humanitarian, but it is definitely like development work that we can do with maps. And this way you also have like a little bit more an impression from myself and what I've been doing. So Nicaragua is also one of those places where you don't have necessarily a map of everything. And we were like in those communities of open source and we were saying hey we want to do something, we want to use this technology to do something that helps us in our city. And Managua, the capital of two million inhabitants, actually hasn't had a map of the public transportation system. Which again seems maybe to be weird but it is actually the case in most cities in the world. I would say that have a public transportation system don't have necessarily data from it. There's also a study by World Bank that says 35% of the world's largest cities have no map and 92% of those from lower and middle income countries. So if you look at Africa, Latin America, big parts of Asia, there is no such information. And we said hey we want to use those tools open street map in that case and we want to get this data and do something out of it. So we took paper, GPS devices, this was the time before a lot of people had smartphones with GPS chipsets in there. And we started collecting the data of the bus routes. We did this in like community building purpose so we trained over 200 people that were interested, everybody could join. A lot of people from universities, professionals, government employees, all kind of people that wanted to learn were welcome and we mapped this together. And we created the first actually public transportation map of whole Central America. And we printed it, we gave it to the people and later on we also created digital tools like an online platform or a mobile application for people. So this is what we can do, right? But what happens if not only one could change that but we could do it all together as we are doing with open source, right? And this with an open street map means that there are one million people that have already contributed to this big database. And it is a global initiative similar to also to Debian that is based on collaboration to create a free and open world map. The map has the data as a free license, it's available for everybody for any purpose and everything that is added is going to be available. It is not the perfect comparison I would say but definitely one of the closest. Open Street Map can be titled like the Wikipedia of Maps, especially because there's a very low entry barrier. So you can like literally as you can click on edit on Wikipedia to add an article or to edit an article to improve the data that is there, information that is there. You can click on edit on open street map and the website would flip around and you get an editor. On this editor you can add points, streets, buildings, you can improve the data and it is really very simple and everybody can do. Oftentimes Open Street Map is seen as this website, right? This open alternative to Google Maps. I wanted actually to do this talk without mentioning this company name but okay. But there's a big difference I think that this is why I think there's no comparison because one is a website where you can do what the company allows you to do and the other one is actually you have the data set and you can do everything with the data set. The planet, you can download the whole planet of Open Street Map that's compressed around 80x, uncompressed around a terabyte. Usually people wouldn't not download the whole planet. You would definitely take a subset of it but it shows that actually it is by now the largest accumulation of open geographic data that we have and this is an extreme value and in a lot of places you can either start from zero or you start with what you have of Open Street Map which is already like a lot of good work. This data can be used in all kind of devices. It can be styled. The maps can look like you want them to look like. You can use them in any kind of application. You can port them to different media, paper, you can have them in mobile phones. You can have special devices in trains or something. I'm a big fan still of printed maps because they're just like accessible. You can use them and also in our case where we work with governments and with humanitarian organizations, the paper map has a big value for planning because of its immediate access. We can do online maps. By the way, this was one of the first online map that I did was the Debcon 12 map. I'm still happy that it's online. We can style them. We can put things on top. We can really dig into the data and we can of course use this data for geographic information systems which is basically doing the professional work with geodata. Now I want you to take not that far from Nicaragua to Haiti. You probably remember the big earthquake in 2010 that hit Haiti pretty strongly. This is the poorest country in the Americas and literally at that moment when foreign help organizations came in but also the local governments, when they wanted to address those issues, they had no geoinformation and they had no way of dealing with that. They had a problem with giving water to the people and to communicate where to get water. This was before everything started on humanitarian work on OpenStreetMap but some people said, hey, I want to help and they used this OpenStreetMap that at that point in time came to a level that it had an editor that it was easy to do before it was much harder. So they used it because they want to help and therefore I have a little visualization for you. On the bottom left you see the days and this is the map of Haiti and you will see in colors the edits that are happening on it. So you can see I think on the 12th it's the earthquake and people start mapping and within a few days there was already like some good information of the city and within less than two weeks there was a decent map of the whole country. So the famous before and after picture, this is the showed progress in only 28 days and this could then be immediately be used without any restrictions and restrictions and was converted and could be used on devices and people were able to start like doing first maps of telling people where they get water. This was the very beginning, there were a lot of issues you perfectly know, right? You have a data set, somebody wants to use it within like three days. But this is where we started and this is where some organizations were founded within OpenStreetMap one that I'm working for is the one in the middle and there are others, there's CartoonG and Le Libre Geograph. They're all doing humanitarian work with OpenStreetMap and we started becoming better. So first we needed to coordinate and we started creating tools that allowed us to work in bigger groups on data collection. We also train organizations, schools and companies to conduct so-called mapping parties or mappetons where people physically meet to learn and map. This time I want you to take to West Africa where in 2014 was this Ebola crisis and this was where we already were kind of professionalized, right? So the international organizations asked us for data but now the pipeline of actually having the data and consuming the data was already set up within Red Cross that was without borders and others. So they had immediate access to this data and they could like say okay, there's something happening in a certain area. They could ask people to map it and then two days later they had it on the ground. To successfully combat Ebola it's required to have the information on a house level because you want to know where some case happens. So you can send your disinfection teams to it and you probably want to do some measuring, some means on the surroundings of this place. For this we created one software, it's called the Tasking Manager, it's a very simple principle. It is basically you define a region that is of your interest and then it would cut everything in like small pieces like these little squares and people can say I'm working on this, I'm locking this so that nobody other works on it and we don't have editing conflicts. So usually people select one place, one task and then they start off with a satellite image, this is how it works. By the way when there's a disaster it's sometimes easier to get satellite images than in regular times I would say. We have like basic coverage but sometimes this is older, if there are disasters sometimes there are more people willing to donate satellite imagery. So based on the satellite imagery people can start drawing manually and this is how it would look like to draw all the buildings in a part of western Africa. If I talk about western Africa I mean Sierra Leone, Liberia and Guinea. And then this is the result of the remote mapping which is then vector data that can be used in all kinds of programs as I explained before. Guacadou in Guinea was one of the critical epidemic centers where it wasn't clear and then they spotted it as one of the places where it's very urgent. And within five days around 250 mappers created this. Now like with the experience and with all the tools that are there and this data was then used by the humanitarian organizations. In total over the whole year of this Ebola crisis there were 2000 volunteers that participated in that and we probably created as a result of this, not just addressing for humanitarian action but also as a result there was like the best map created of western Africa. Which by the way nowadays is used of course for other purposes and that's great. Here just like a small logo introduction of like what are the organizations we are working with, there are more. This is just like the ones that are more related to the examples that I'm giving here. And with those organizations we were like seeing hey but there are still so many places in the world where we don't have geographic information so how do we do that? This is not energy consumption of the world but it looks a little bit like that actually. But this means where we have data in open street map and if you look at like India, China where more than 3 billion people live, there should be more data on it, same for parts of Latin America and Africa of course. So we created this initiative called Missing Maps where we want to put the most vulnerable places on the map. Why we say the most vulnerable places because surprisingly there's a correlation that often times the places that don't have geographic information are also the ones that are hit, versus by disasters. Missing Maps is definitely an armchair project so people sit, meet in Mapitans in Austria like you can see here or sitting in the comfort of their home to contribute. But of course we are doing Mapitans everywhere in the world. And we also go to places where people have and we learn together and we also go to places where people are learning with us too. So we even did the Mapitan in the White House, not with this president of course. And I think it's particularly interesting because everybody who gets into open street map has to learn. And seeing that these organizations and governments are catching up with learning what means open is great to see. I will come back to that later. This is in German but it's a placeholder for this video that I wanted to show you which is from Doctors Without Borders and gives you like a good impression of how the humanitarian organizations now perceive the opportunities open street map can give them. Okay then we realized of course we knew before but it is really important to highlight that remote mapping is of course possible but it is not the most suitable form of mapping. It's from satellite imagery you can see houses, you can see roads but you don't know whether this imagery is from today or from yesterday and if the house is still there and of course you don't know the names, you don't know the function, the purpose of a building, whether there's a pharmacy or there's a casino in there. So from a data perspective you need definitely local knowledge but for me knowledge is much more important in a way of I want to differentiate it between information and knowledge while information is the artifact and knowledge is something that can only live in the head of a person because it's like when the information is processed by yourself. And this is really important for open street map because every place in the world is different and also how we started we didn't start like with one smart guy or girl that was saying hey we have this data structure and now let's stick to it we are like actually moving as we are going and the local people are by default when we talk about maps they are the experts. And this is why it's so important to let them become part of the whole movement. To explain this a little bit more I want to give you a little bit an introduction of how the data structure works in open street maps so there are only three types of data, there are points, lines and relations and with this we can actually build up polygons, lines that are related are a polygon we can do of course all streets and we can put points on it. Those are classified with a simple tag value pair or with key value pairs. So for example here we have name and a name in highway classification residential. This is a very volatile data structure because there are no strict rules you can basically put everything in there and people put everything in there, literally. We have a wiki so we organize ourselves of how we use and how we formalize our data. We do this through the wiki, we do this together, we fight a lot of course. We don't have a last instance that takes a decision. So we are there much flatter than for example even Wikipedia because really everybody can add it and it is that simple and there's no privileged role of being there for a longer time or something like that. We do have a police, we have a data working group that's like three, four, five people. They have almost nothing to do and we have a lot of people with very curious eyes looking what people are doing so this is how we defend ourselves. And of course we have bots that look also for malignness edits. So coming a little bit back on this local knowledge, I think here it starts becoming clear why it is not just about the data, it is about the people that are taking part in this because we are constantly defining the data structure and it's changing and it's different in different countries maybe. And it's different even in the same country if two people have a different idea on how to do it in their specific region for a specific topic. And we have to constantly deal out how we move forward. And in this context it is just like very interesting to see what are the particularities of OpenStreetMap in comparison to other open source projects. And I think here there's really to see the low barrier that I mentioned before of the full open contribution cycle. So you can enter, everybody can enter, you don't have to be a full programmer. People that are just curious can enter, click edit, they have, there's a map, they put one point more on it and they saw hey, there was something, I contributed and I have something bigger that I got out of it. This you can explain in one minute to almost everybody with some visuals. And by the way this was also the reason why Nicaragua that became popular because we really had the chance to reach out to much more people outside of the bubble of our technical people. Because it is clear, it is very, very obvious to people when it comes to visuals and when it comes to maps. On the other side I think this is, I recently got inspired by the Open Knowledge Foundation that were using those three components as successful components of openness. Of course the open data, the data literacy that I think is oftentimes not enough pushed ahead and this is what I was also meaning when I say knowledge. And the tool, the free software that allows you to actually deal with the artifact. And I think one particularity of OpenStreetMap is that those three are somehow equal balanced. And this also of course plays into the possibility that more people are getting into it and understand the cycle because all three of them are there and get into each other. You could not take one piece out. Not at all. And this is true, I mentioned it before on the World Bank, on the White House map aton that we have really got into a level of collaboration that is huge. So we have a lot of craft mappers, people that want to do it. Almost all Silicon Valley now uses OpenStreetMap besides the big G because they already have the monopoly on their stuff. Development agencies, UN organizations, humanitarian organizations, they all went into that and they all had to learn through that how to use open source and open data. And this is nice to see that this is like happening for all, not just for the people that we see, that get in the first touch with technology but people that are very huge but will comply with these rules. Let's do quickly this last example which is in Tanzania where people are having a lot of problems with floodings and this has an effect on the health of the people. So we were using another tool which is Open AerialMap together with Open DroneMap that allows you to actually take yourself airborne imagery, upload it here under an open license and it can be immediately used. And again we created a wonderful map of the capital of Tanzania together with students, a very local approach to that. And I would say this is a very nice success story because then they founded their own organization and they are now spreading the word and the knowledge around OpenStreetMap in neighbor countries. The data itself is used by the city administration for planning now so a lot of work has been doing and then getting closer to the authorities here and to improve the situation which is of course easier to do if you have a map to show what's going on. Quickly, you don't have to download the planet, you can do queries, there's an API for that, I think I don't have to go too deep into that if you don't like APIs which I don't expect in this room. There's also an export tool where you can do this with clicking, important for some organizations. And to finish just some last examples of how this data can be used, here doctors with outborders took the buildings from OpenStreetMap based on this data they estimated the population in different areas and they knew how big their bag for vaccines needs to be if they go to certain places. In the Philippines, which is highly exposed to tsunamis, we developed Inasef which is a quantum GIS plug-in where you can run risk scenarios so you can use the data from OpenStreetMap. You can say, in this case, you combine it with a digital elevation model and you say, okay, if the water comes up 10 meters, what happens? You can do this for planning and you can do this in the moment. And of course, in those times, we cannot do a talk without mentioning deep learning. And we do some experience with it. We have to be very honest of what can be done. One thing that can be done is you can try to estimate what is on the images and then you can compare it with OpenStreetMap data and then you maybe have an idea of where there's more work that needs to be done and where not. This is so far how we have been going. OSM is a big ecosystem. A lot of things that are not called OSM are part of it. So also a lot of work is happening in Debian. We could not survive without Debian, so thanks a lot also for all this good work. And if you are interested, you can always go to tasks.hudOSM.org and start mapping. There's always good work needed. There are links on how to map. I'm happy to share the slides also afterwards. And there's a mailing list if you are more into that. Thank you very much. I have two questions. If I got the process correctly, the first step if you are mapping a new area is to take satellite images and then later the locals augment their data by adding street numbers or street names or something. The first question is how do we know that the coordinates on the satellite images are actually correct so the image is not displaced by some meters because it's a slanted area. And the other question is once you've mapped the area, how do you keep that data up to date? New houses popping up, so are you looking at the satellite images again or does that work? Two very good questions because to both I don't have a good answer. The first one is the alignment or misplacement of satellite images. We have to deal with this all the time. So we have usually several sources and we see that they are not well aligned. We have possibilities of aligning them also in a community effort so you could like on certain areas say this is an offset like this, this is an offset like that. And then displace the whole map? The whole tile. Like they are working in tiles. Yes, it would correct that. Of course this is why we need people on the ground, we need GPX data on like GPS exported data on that we can combine so we can see this line is the street and then if it's the satellite images there we start like moving it there but I have to say it never fits 100%. It's always an approximation. And sometimes it is also like it is off for some meters. And the second question. Keeping the data up to date. Keeping the data up to date. We are still there and we try to get all the data but we are already anticipating that keeping it up to date is actually more work than building it because if you have an empty paper it's easy to write something on it or to fill it but if it's already full it's more difficult to find which is not good anymore. I would say the only solution for this is actually local communities that have an interest because if you are living in a place you are verifying constantly and you have an interest in keeping the data up to date. I started doing OSM a long time ago and recently did some hot mapping which was great fun. Terribly, terribly addictive. It's really dangerous. And we did it at work so you go excellent I can call this work. Brilliant. So I did like two weeks of nothing else. But the biggest problem I found was the dating of different aerial views. So in the middle of West Africa you've got different satellite vintages and I found it very hard to work out which was newer when they're not the same. And does that metadata exist? Is there an easy way to find out or is it just hard? Not really. It's just hard. We would like to have better... There's one problem on having suitable metadata on satellite imagery which is sometimes still hard because the providers are not doing that in a way that we would like this to have. Of course, we would like to have much more open imagery but it's very expensive. So yeah, it is reading, asking, looking. Sometimes even the providers have like you scroll a little bit to the side and it's another year then it is there. It is tough. Yeah, we definitely don't have a perfect situation on sources for satellite imagery. If nobody has any questions anymore, let's thank Felix again. Thank you.