 So thank you very much for the kind introduction. It's a pleasure being here in Oslo, and I'm honored to be able to talk to you this morning. As mentioned, my name is Anita Grazer, and I'm from Austria. I work in Vienna. I'm a researcher at the Austrian Institute of Technology, where I focus on transportation research. So I'm looking at people moving and trying to find out what they're trying to do. Besides that, in my free time, I'm very involved in the open-source GIS community, most notably in the QGIS project where I'm part of the project steering committee. That is the group that tries to organize all the boring stuff behind the project. And then I answer questions on state overflow and try to do my own side projects, which you've probably seen around the internet. In addition to that, I'm also currently serving on the board of directors of OSGU. That's the big umbrella organization that hosted also the phosphor G in Bonn, and then it's home to many other open-source GIS projects. But I will talk about that a bit later in detail. So the whole story behind how I got involved into open-source GIS was really I started out as an intern in a research company. It was summer, everyone was on holiday, and so was the GIS person, which was great because no one knew where this ominous GIS license was and how I could get access to it. So I was there looking forward to a month of internship without the GIS. I had to sit down and try to find a different solution and Google around and try a few of the GIS tools. A few months later, the QGIS developers met for the second time to have their developer meeting and it was in Vienna, so I joined them and I've been stuck since. It's really the community and the speed of the project that got me and I'm still staying there. I already mentioned I'm working at the Austrian Institute of Technology. We are a non-university research institute, mostly in Vienna, which is owned by the Republic of Austria and the Federation of Austrian Industries around half-half. And we have a lot of people, there's over a thousand people working there on five main fields, which is energy, mobility, that's where I'm working at, safety and security, health and environment, and innovation systems. The small area you can see there, that's where I'm currently sitting, and I've been told when I come back I'll get to move a few months over. But overall it's a very nice view right there in Vienna. My group is called Dynamic Transportation Systems. We do everything, research on everything that moves on transportation infrastructure. So pedestrian trends and how they move in events or inside train stations. And we also look into how to collect data about movement, which is not as simple as we sometimes think, particularly if you want very reliable data. And then when we have collected data we try to understand it and we try to model it and make forecast models. And then there's also a big group, which is getting more and more important for our group, which is the dynamic route planning, optimizations and logistics, which has a lot of talented people in it. So the AIT pays the bills. And the QGIS project, that's really where my enthusiasm goes. We use it a lot at the AIT, we don't have any other GIS licenses currently in my group. The QGIS project was founded in 2002, so it's already over 10 years old. It all started in Alaska with one single guy, it was Gary Sherman, who sat there and he had this new thing. He had this post-GIS database with all the data that he could finally work with, but he had no way of looking at the data. So his whole motivation was he finally wanted to have a look at the data so he developed a viewer and that's the project that he called QGIS or back then Quantum GIS. Quite soon other people around the world joined him when he had released the code and it started rolling rather slowly in the beginning. You can tell that from having seven years between the first release and then going to 1.0 in 2009. I myself had joined it around 2007 when we had the version 0.7. From then it really took off, so we had four more years until 2.0 and I will talk a bit later about the plans for 3.0, which is around the corner. So we've been constantly growing and to channel all this work and to meet because there are so many developers all around the world from New Zealand to Alaska. We try to meet twice a year with all the people who can make it at developer meetings and since last year we also have a yearly user conference which was organized in Denmark last year and in Spain this year and it's a really great event to get both users and developers together in one place to discuss what's been going on the last year and what are the plans for the future. This is a picture that we took from this year's user conference in Girona, which was a real pleasure to attend. So I've already mentioned that the project has been growing continuously over the last years. Here you can see the number of contributors per month. Those are the contributors that can directly commit code into the repository. So this is the last underestimation of the number of people who work on the project. I will get later into how many people are actually involved in running a project like this, but even in the committers you can see we have a nice increase over the last 14 years already. And I think these numbers, these estimates that we get here from OpenHelp, those are really impressive considering how much work has already flown into this project and how mature it is. I already mentioned OSGU for a moment in the beginning. QGS is an OSGU project. It's a graduated OSGU project. That means it has a certain organizational structure which is a project steering committee and there is ways to contribute to the project and it's not sealed in but rather open to the public and to new inputs. OSGU is not for profit which is registered in the US and its goal is to foster open source uptake to help the project to support them if they have any trouble legally or in other ways. And it drives outreach projects mostly, for example the conferences but also GEO for role which is their program for academic, for teaching materials to reach out to universities and make them adopt open source in parallel to the commercial offerings that most universities currently have. Furthermore you will also be familiar with OSGU Life which is a tool also for academics but in general for outreach it's a collection of all the programs that are within OSGU and it's basically your first starting point. If you've never tried it out, get an OSGU Life, download a USB stick from one of the conferences, you just plug it into your machine, you boot it and you can try all the software. It's pre-installed, no hassle there. It runs on a lube onto operating system and it comes with test data so you can just give it a try without messing up any machine. Coming back to QGIS. QGIS is an open source project but knowing that is not enough to understand the nature of how it works and how it operates because there's many times of different open source projects. And there's this one distinction that is quite popular between the cathedral style open source projects and the bazaar style open source projects and I think QGIS is very far on the bazaar side of these things. Basically the main difference is in a cathedral there's architects, maybe just one architect and he or they they try to control the whole project and they sit and they do until they are happy and then they release the source code. So it's open source but while they are doing and while they are filing at the different pieces, bits and pieces you don't get to see what's going on. On the other side in QGIS you have this completely open system of the code being always visible, everyone can branch it, everyone can work on different ends and pieces and that way we collect a lot of different inputs. Sometimes you have inputs, features that attract the same crowd, so you have different people working on different issues and sometimes on the same issues and sometimes there's duplication of work but the big advantage is when people get together and find out, collaborate and that's why the bazaar is the model of choice for us and it has served us really well. In this paper by Raymond he says so nicely open source will triumph and not because it's morally maybe superior to say we want to share things, I mean if we get real who cares really, we all have to eat, we all have to pay rent so morals are nice but it comes down to survival first. But the point is and why he thinks open source is a good model for development is that the commercial, the proprietary world just cannot win the arms race of all these people are contributing to open source projects because they simply cannot put in the same manpower and skill manpower to solve the issues. I mentioned before that this number of developers that we saw in the graph is vastly underestimating the number of people that are involved in an open source project, especially an open source project that is as huge as QGIS. So I have here this graphic of the QGIS team as I call it which I like to show because it's really important for me to mention that the developers are an important part but they are just one part of a big puzzle in the whole thing. And then there's many other groups which are equally important and keep the whole thing running and making it usable for everyone around the world. For example we have the translators who make sure that QGIS is available for local users all around the world in I think over 70 languages by now. And they do an awesome job because they have to always keep up with all the new features that are introduced. Then we have of course the infrastructure managers who take care that the downloads are available when you need it, that the backtracking is available, that you can have all the infrastructure accessible for both the developers and the users. Then we have the huge group of documentation writers and they have a really hard time because as you might know we are releasing QGIS three times a year. So there's a lot of documentation that's always to write and also a lot of new features that want to be described. So we are really proud of our documentation team and they are always looking for more volunteers. Then what I've put here is community managers. Those are the people that are looking at the users and trying to support the users on the mailing lists, on JS Stack Exchange which was already mentioned today and in other places on local forums which are not in English but there's a lot of French forums and Spanish forums for example. So they try to keep all the people together, they collect feedback from the users and they bring it back to the project which is a really important task as well. Then you also need designers in your project particularly to keep the website model looking to half a good user interface and to try to do marketing outreach which doesn't look like it's been done at night by some developer who would rather have written codes than to make this kind of work. And then what's also very important to me, I consider the users as part of a team of an open source project. The users are super important because you shouldn't underestimate that developers involved in open source projects quite often aren't directly also the users of the open source project. Quite many of them are great at developing but they are not power users. They never use any of the cheer processings functions themselves so they probably will not notice if there's a workflow which is not ideal. That's why we need this feedback from the users which tell us okay this could be done much better the current workflow you always get stuck here that's made more complicated than it has to be. This feedback from the users is essential and it's the job of the user who is not the customer of the open source project but part of the team to also contribute. If you want to be a customer that's fine but you're the customer of the commercial support provider who works with the project. You can never be the customer of the open source project it's not how it works you're part of the community. Also part of the community and super important for us are our donors and sponsors. Without them it wouldn't be possible to have these developer meetings. It wouldn't be possible to do many of the other things that the project is trying to do like increase code quality by having regular practices and also run other outreach programs. And of the bottom of it all I put the project steering committee. It's the group of people who takes care of signing all the necessary contracts. We've trademarked the brand and the logo of QGIS so just in case to avoid any trouble. We have all the contracts for the servers and the domains and all this stuff needs to be taken care of. It's not very glamorous not very exciting but that's what the project steering committee is for. We've also been working on forming a new legal entity for QGIS because so far all these contracts were signed by individual members of the project steering committee. And after 14 years of existence we thought it's time to professionalize that a bit and to have a legal entity who can actually sign things. So there is a new organization association registered in Switzerland called QGIS.org which we are currently setting up and then we will have an official voting for the board of directors of QGIS.org. And if you want to vote voting members of the organization are nominated members of QGIS user groups. So we already have 13 user groups official ones that registered with the project and each of them can send one voting member to the association. And in addition to these voting members from the users we have an equal number of voting members from the developers to keep those things in balance. Because we really want to be a developer group organization. Because as an even number we have one other voting member which is from OSGU who can break a tie in case developers and users just cannot agree on a thing. I don't see it coming but just in case that one is there as well. So it would be great if for example from Norway I think we don't currently have a user group from Norway. If you form one you also get one vote in the association and you can decide on things like how we spend our money. How we spend our money is a public thing. You can find this financial report for 2014 and 2015 on the website. And it shows you how much we've made mostly in donations and then sponsoring. And what we spend it on. And last year we spent a lot on bug fixing. So there's one bug fix round before every release. It's organized in a way that developers who have time to do the work they apply. And so far the PSC has decided whom to pick. We always had a set amount in our budget and we went along with that. We also have spent some money on the developer meeting expenses. That can be travel support that is mostly for the food at the conference. And we try to keep costs very low by having a venue which is usually at the university. So we don't have to pay for those things. And then there are some other expenses like the servers and all this other thing which you have to take care of. And this budget and how the money is divided is one of the things that the voting members will decide on. We also have some new ideas to not just invest in bug fixing. We also want to invest more in documentation both for the users and for the developers. So we have started a so-called grant program where we ask people to apply with their ideas for what could be done to improve the QGS project. And we hope we get a lot of submissions which deal with that kind of issue of how to improve documentation, about how to improve our infrastructure, maybe some outreach things that we then can support with some money. Speaking of documentation, there's tons of documentation available online and this has been like that for many years. But what really did take off in the last two years, three years is the books. And I have to say I think the books really showed that they're serious. And they had another big effect with the books and the QGS too. We finally saw a much bigger uptake of QGIS in education, in academia. With universities finally realizing, okay, this is not going away. They are serious about this thing. You can even get something that you can put in the university library so I can finally go to my superiors and I can say I want to offer a course for this software as well. And it's an important asset for our students to know how it works. So there is like a lot of courses that you can find online. I'm currently also teaching at the University of Salzburg which has a distance learning master's in geographic information science. And we're also using one of those books. Besides that, there are self-study materials available from mostly US American universities. Like the GEO Academy curriculum which is a good starter if you want to get an introduction into the whole topic of QGIS. Because they really made it step by step and you can follow along really easily and there's no need for an instructor in the way that they organized these learning materials. Of course, all these learning materials, they will never cover everything. We need support channels and there are a lot of different support channels and ways to get into contact with the project. The classic thing and I know it's frowned upon by particularly the younger crowd are the mailing lists. Yes, they are old school-ish but they work and the developers are used to working on those. But because we know that many people rather want to have a forum and something that's a bit more interactive, we also support on GIS.StackExchange which is a great platform for asking questions and finding answers because it's really indexed very well by search engines. So we used to have a forum before GIS.StackExchange started but it was really a problem because if you were googling for the questions you couldn't find them on the forum. And I think now with StackExchange that has much improved and I would recommend everyone to use it. I already mentioned the user groups. They also provide local support for their members because they often have to deal with very local issues with specific datasets like you also have in Norway, you have your own dataset standards and that's the same case in many other countries. So they deal with that and they also lower the language barrier because the official mailing lists, they're all in English and some people have trouble describing their issues. So they rather go to their local mailing lists or they have local forums where they can interact with the community that's in their country or in their language area. And of course there's also commercial support and I say of course these people have been working with QGIS for a long time and it's a myth that you cannot get a support for open source projects. The ones that I've listed here are just the ones which have core developers on their stuff. So they have the people who can actually commit code directly into the QGIS project. But there's many more which are not listed here who have developers but the developers have to submit pull requests on GitHub to get the code into the project. That doesn't mean that you cannot get it in, it's just one extra step because we want to ensure the quality and we do laugh by having the core developers check it because they've been working on the code base for a very long time and they know the ins and outs and can estimate what effect it has if you change something in a certain place. If anyone is interested in joining this group of core developers, it's not a closed group, anyone can basically apply but you apply with your work. So before you can join it, you have to demonstrate your ability and your willingness to apply good code to the project and after a while you can get invited to join the core committers who have some certain powers but also responsibilities in the project. I mentioned before, it's getting time, it's getting exciting, we're moving towards QGIS 3.0. Why do we have to move? We have to move, it's not something that we decided on, it's not an our timeline. We have to move because many operating systems are moving from Python 2 to Python 3 and also from QT4 to QT5 and we want to stay up to date with these dependencies. They have a lot of advantages and we hope that they bring us advantages and not just work. So we will have QGIS 3 with all those new libraries. We think the first version of 3.0 should be out somewhere in spring. We will have to evaluate the timeline and see if that's doable around after Christmas, at least that's the plan right now. I would recommend, particularly if you're interested in the plugin development, that you try 3.0 because that's important to note. All plugins for QGIS 2 will be broken when you try them in QGIS 3 because Python 2 and Python 3 are just not the same thing. So everything will have to be updated to Python 3 and the new QGIS API. We're planning to have guides for plugin developers which show what are the usual things that you have to change. Of course a list of API changes to make it easier for the plugin developers to get to the new version. But still you will have to touch your plugins if you have any. That has the advantage that we will probably be able to narrow down the number of plugins that we have. Currently there's a thousand plugins in the plugin repository. I think no one has ever tried them all. And it's going to be really exciting to see how many of those will be updated to QGIS 3. And I also hope that we can consolidate a couple of those plugins which do very similar things into one better improved version which will then ship for QGIS 3. At least that's my hope. Other hopes that we have are bound to QT 5 which is supposed to have better support for mobile devices and for high-TPI screens which you probably have on your new notebooks. So supposed to look nicer then. And there's also other advantages that we hope for, for example, like the SVG export which is done with QT. That's hopefully improved in QT 5. So let's go from the organizational and from what are the plans for the future to what has been the latest improvement, what are the latest features that QGIS has gained. There's always a really, really long list so there is no way to cover them all in a talk. And you can find this list always in our so-called visual changelogs. They're visual because we try to add pictures and animations to show exactly what it is. So you can get a quick overview of what's been going on in the project in the last four months since the previous release. And it's always really fun to scroll through it. You also have a mention of who sponsored the features. If it's a sponsored feature that was contracted by a company. So you can see who's involved in the project, who has an interest and who are the developers as well. So if you have a certain task that you want to have implemented, check through the changelog and see who's been working on similar things, who knows their stuff. So let's dive into a few of those features that have been landing in the previous year. The first one seems 2.8 since the last long-term release version. One thing that I really like a lot is that we finally have a start screen, not just a blank map, but an actual start screen which shows you the last few projects that you've been working on. Which is really nice, especially if you have a lot of things going on at the same time, chuckling between them, to just with one click find exactly the project that you've closed the previous week or two weeks ago. That's a really nice way of having a look at the projects. We also have a new feature in the map view. If you have trouble seeing it, for example, because you have a high DPI screen, there is now a magnifier tool in the map view which you can use to increase the visibility just for working on the map. And you can also rotate the map now in this view, which is a nice additional thing. One big overhaul, and I don't know in Norway how the disaster looks like, but this has been driven a lot by this Swiss disaster, is the new geometry model in QGIS, which supports not just simple features, so always straight lines between points, but arcs and circles and all these things that you can design in CAD software. So you have direct support for a circle. This circle just has one dot and then a size, and it's more in that way rather than many, many different nodes that form something that's kind of a circle, but not really. So in the back end this has all been implemented, and if you have a storage system like PostGIS, which can store these arcs, you have the whole workflow, and you don't need to go back to simple features anymore. So maybe that's also interesting for the Norwegian disaster or engineering data sets that you have. There's also, on the slides, there's also support now for all the 3D data, so if you have a shapefile or any other data which has 3D coordinates, they are now maintained throughout the processes, which wasn't the case before. If you loaded something in 3D, QGIS just silently dropped everything in 3D or M values, and when you exported stuff it always ended up as 2D because they lost all that information. That should be solved now as well. There's also tons of new functions being added to the expressions like all the time. I've picked some out which are particularly interesting for photography, so that's something I like to work with, I like to explore what's possible, push to the boundaries of the styling functions. So you now have a way to, much easier, set different parts of the color to, for example, make it go from very light to very dark with the color part of setting. There's also an even more convenient function which just says darker, so you give it a color and you say make it a bit darker than that, and magic ensues you don't have to remember any RGB codes or other stuff that you do in other tools. There's also a reverse, so you can flip around lines for example if you want to draw arrows, but the roads kind of have an attribute which says is it in the direction that it's digitized or in the other direction, so you can flip stuff around. You can even do closest point computations now in the expression builder. So you give it two features and it finds out it's the closest point on the other feature and you can mark that there. And one really cool thing that you can aggregate in the expressions. So you can finally say, give me the maximum value of all the population entries in this layer and, for example, calculate the population share of this feature relative to the whole population. That's something really new because before that you could only compare it to other attributes of the same feature but not to the whole table and that makes many workflows just so much easier. Another workflow improvement thing is in exporting data, you now have fine control about which attributes you want to export and how you want to call them. So, for example, if you want to export to a text file, you get to choose from all the columns just exactly which do you want to use and if you want to rename them, that's also possible here. So, if you've ever worked with like joint data which ended up with 30, 40, 50 columns and you had to then open it in some other editor and remove columns, that's over now. You can all do this step directly when exporting from QGIS in this tool. Another thing that's really nice to work with when you get used to it and which a lot of people wanted is conditional formatting. Probably you know that from Excel. So, you say, in this column, if a value is under 5, make it red and if it's over 10, make it green. And you can do the same thing in QGIS in the attribute table. We have the same conditional formatting. Maybe it's even more powerful. So, you just define your conditions, the colors, and you can see how it translates to the attribute table and how everything is highlighted and you can sort by it. It's a really nice and nifty tool. That is not obvious to everyone because it's kind of hidden, so try it out. It's back there. Hit another button in the top right corner. There are also many more improvements in the attribute table. You can finally have a context menu. If you click on the name of the attribute in the attribute table, you can hide columns, you can change the width of the column there, and you can sort, which was really in high demand for a long time. You finally have a way in the attribute table to just sort it visually, not in the data set behind it. It's a little bit slow because you're changing the data set. It just changes the display order. And you can also add actions to the attribute table. So, a button right next to each line in the row in the attribute table to do a specific task. For example, to open a file where the path is stored in the attribute table and you want to see that file in front of you. And all these workflow things, there's tons of new improvements for visualizing the data. One thing that I really like a lot and that I want to use more, hopefully, in where it's appropriate, is the 2.5D renderer that you can see here. So, it's basically a block model that you can extrude the height of a building depending on the fixed value or depending on an attribute. And you can also style it and do all kinds of other fun things with the extruded areas. So, this is very useful for these building visualizations but we've also seen a lot of statistical maps, for example, where you have a certain value for a country or a district and you just scale it and you make the surface model out of it. We also have a new renderer for arrows. If you've ever tried to make some arrows in the past in QChill, yes, it kind of worked, but it was a bit hacky. Now you have arrows and you can even have curved arrows and those are really neat. You just need three points to define the arrow and it automatically calculates a nice, smooth curve for them and they can be highly customizable because you can apply every kind of polygon style to these arrows and if you've ever played around with QChill and styling, you'll know that there is billions of options in the polygon renderers. Another thing which popped up like three months ago was a guy on Twitter who said he had written a live hillshade renderer. So it was just a small Python script that he ran in the Python console on a digital elevation model and without actually writing a new hillshade file it would just display the original digital elevation model as if it was a hillshade. And this idea was picked up really fast by Nathan Woodrow, one of our Australian developers and now we have it in the main application. So the script anymore is actually part of QGS Core and you can just for any digital elevation change the display to hillshade renderer and it will create the hillshade live without writing a new file and you can change all the parameters if you like them and it's also live updating. This live updating you definitely have to try it's new in 2.16, so get that version and you will have what formerly was in the layer properties in a dedicated window is now on the right-hand side of the map and whenever you change something here it's automatically and immediately updated in the map so you have much faster time much faster response to any of your style changes to any of your label changes you see immediately what goes on if you change this one parameter it's so much easier to experiment and to fine-tune your map with that and I think that will be one of the tools that really improves maps because it's less tedious than always having to go somewhere opening two windows and change the color to a bit darker or a bit brighter so now this is just one click away and you have live updates and I think that will encourage people to make much better maps and increase the speed at which we can produce these better maps so I already mentioned you have the same capabilities for labeling so it's right here besides the map and whatever you change here is reflected immediately if you are worried that maybe that would be a bit slow in some cases it can be there's also a tiny button here where you can turn off the auto update but I've worked with reasonably big files and usually it's not an issue so I just keep it on for now and enjoy that I don't have to click apply anymore at close windows you can also open the attribute table at the same time I mean that before you couldn't do you couldn't have a look at the attribute table and the layer properties at the same time because the layer properties was always blocking the rest of the interface it's all gone away with this little addition which I think is just wonderful of course there's also improvements to the algorithms themselves for example the labeling algorithms there is now a better automatic placement so if you should first put it on the upper right corner and then in the lower left corner of the symbol you have also options now for label obstacles so you can say never put a label over this feature although labels have to be put somewhere else don't cover this one up that's one of the new things that's supported and we also have this really nice thing of that the labeling algorithms now considers the symbol sizes so you can tell here all the symbols have a different size but the label is always nicely relative at the same distance to the border of the symbol which was formally not possible automatically you had to do it with data-defined settings but now it's much simpler to achieve another thing that happened on social media which started on social media and that's what I enjoy so much about these fast development environments that we have in QJS is this gradient editor thing you probably some of you might know the old QJS label gradient editor which was just okay you have to can define two colors and if you want to have more colors you have to type in at which position of the gradient you want which color and it was really not very interactive and user-friendly and on Twitter we discussed that one day and it just took two days I think until it landed in the nightly builds and it's not just here that you have this gradient editor like you're used to from graphics design software but we also have this very nice new edition which is rather novel I think in most fields is where you can control actually different aspects of the gradient much better you can control for lightness we can see here in the demo when I move this up and down the color gets lighter and darker and you can create gradients that ensure that for example it always goes from dark to light so if you print it in black and white it will also make sense compared to your usual gradients which quite often screw up if you print them in black and white or you can also make like this gradient which is brightest in the middle and then gets darker to both sides and you have really fine control here which I also think will lead to improved maps and much better visualizations I hope my totally subjective action features showed you how exciting it is all the new things that came in please have a look at the visual change box as well because I certainly left something out that would have been very interesting to you as well I want to use this opportunity to also encourage you to get involved in a project there are many ways also for non-developers to get involved in this project you should definitely report issues if you find them let us know about them because otherwise we cannot do anything about them please help us translate and document our features we just looked up the Norwegian translation yesterday and it was somewhere at 80% so there is still room someone has started it there is still room to improve then we can always use bit more funding to squash more bugs to make our infrastructure better if you already know something and if you are interested in learning more try to engage and to answer user questions it's a great way of learning about the commonly faced issues and how to deal with them it will also help you within your organization because really 90% of the problems are always the same of the same issue and if you are a programmer of course it would be wonderful if you were to help us find and fix issues and help developing as well so my title was QGIS Community Power GIS and I think it is so great because it empowers communities all over the world and I hope it empowers you as well thank you