 Thank you first of all for inviting me tonight here in the British Embassy, it's actually my first time here so I really enjoyed it here, it's a beautiful building. And I want to start by of course introducing myself, I'm Isis Anfeldt, I'm leading the news lab at Google for Germany, Switzerland and Austria. It's the past four years and I joined as it was mentioned before, I joined a month ago, so it's a whole new world for me as well. But for the past four years I was working at Twitter and being responsible for news and government relationships and the strategic relationship building there. And when I thought about the topic of tonight, making sense of big data, I actually realised that even before joining Google and before joining Twitter, I was working or I was trying to make sense of big data myself. That was 2009 and I was writing my master's thesis in political science and I was analysing the social signals during the 2009 Chancellor's debate. And I was focused on tweets at that time because Twitter and politics played a big role. And I think I analysed around 40,000 tweets manually, which took a long time. And now I know that it takes maybe a few seconds. So a lot has changed since 2009 and I want to talk today about how to make sense of big data from a really journalistic point of view. So a lot of things have changed in those years. But something that doesn't have changed is that the ticket doesn't work. No. Okay, so that hasn't changed. It's Google's mission and it's to organise the world's information and make it universally accessible and useful. And if we examine that mission in more detail, it's hard to think about a more important source than or a more important source of information than quality news content. And so we at Google News Lab and it's a really new team and new effort at Google. We at Google News Lab think about that in a way that we want to collaborate with journalists, academics and entrepreneurs, media entrepreneurs, to drive innovation in media. And so the ultimate goal of the News Lab is first of course to collaborate with those players, journalists, academics, media entrepreneurs, the new voices in media. And we do that by three or in three different ways. The first one is tools. We want to make sure that journalists and academics have the best tools at hands in order to make sense of big data or to really work or really enhance their storytelling. Second is programmes. We want to collaborate with the new voices in media and we work for example with the European Journalism Centre, Hex and Hackers and many other international, not only international but also European partners to create this level of innovation and to hopefully create new products or new ideas around storytelling. And the third part is of course data. Our goal is to provide journalists and academics and the new voices in media with the best of Google data and to help them create immersive, really interesting, really innovative stories. And tonight I'm going to focus of course on the data bit from a journalistic point of view. And I think the most successful or not successful but the most important data set that Google has is search. So just one number, 3 billion searches every day, over 100 billion searches every month. And that can be a really interesting source for journalists and academics. And because we have this new really collaborative approach at the news lab, at the news lab, we think that creating a more informed world, more informed society, journalists and technologists have to work together. We came up with a really broad curriculum at the moment but we listened to a lot of journalists. We visited many, many newsrooms over the past year and we figured out that not every journalist and not every academic has access to those tools and has access to Google Trends data for example. And so that's our biggest effort that we now trying to follow up on. Google Trends is the starting point for journalists and academics. And Google Trends or the data behind Google Trends has been in the past okay, I would say. So and also we talked to a lot of journalists, we talked, we spent a lot of time in newsrooms and journalists and editors told us every single time that the Trends data is not real time. I'm going to show you how it looked like. So that was Trends before this summer. You looked for, for example, Zepp Blutter and you got this graph. It's not really compelling and you can't really tell the story behind it or create a story behind it. And so the team, the Trends team at Google and within the Google News Lab thought a lot about it. How can, how can we create or make Google Trends a better data source for journalists and academics? And this is what Google Trends looks today. It's much more real time and it shows you the peaks of interest. So the peaks of people searching in that example for Zepp Blutter. And this is only a really basic search career that you do. And in the next couple of minutes I'm going to show you a couple of examples about how journalists in collaboration with us or without even our support made sense of Big Data, made sense of Google Data. And what we also incorporated in the Google Trends setup was Trending Stories. So Trending Stories feed that showed you and which is based on not only Google Search but also on other Google products like YouTube or others. And Trending Stories shows you in real time what is trending on Google. So it's the longest, I would say the longest feed of breaking news stories that can at least point you to a topic or to a story. Of course, the journalistic investigative work follows up on that. And it also shows doing, for example, breaking news events or moments like the US election that is coming up next year. It shows you different ways on how to use and how to analyze Google Data on what you can see on this election 2016 Google Searches site. So what Google Data can really tell you are different things. It can show you what people really care about because they search for different topics in real time and they're pretty honest about it. So and I'm going to start with a really simple data visualization based on Google Trends. Even though I'm not sure how many Halloween enthusiasts are still in the room but I'm still going to show it. So most of you know last weekend, it was Halloween, SET which is a new online site launched by Site Online recently, looked at Google Data and wanted to ask who's interested in Halloween in Germany and where should we look for party tips and how or who looked for party tips. It's a really basic question that lies behind this data visualization. But beautifully visualized in that way. And I didn't show that because I think it doesn't really fit into my whole run up of the slides but we also looked for costumes, dark costumes for Halloween. So even you can search for a dark costume at Halloween. Very simple and quite funny. But what you can also do with Google Trends data is looking at sentiment and looking how people react to certain events, breaking news stories that drive your market or drive another story that is following up. So BBC News in that example looked at the most searched for leadership candidates of Labour. And you can see on the map not only one instant and one snapshot of it but you can actually look at this data from seven days or 24 days or a month or over a year. So you can see the changes in interest in political candidates or in that case the Labour Party, Labour leadership candidates. And whenever there is a global event happening people either join the conversations on social media platforms like Twitter or any other platform or they go to Google and search for more information. So when you know what is happening, who's involved or what's the background story of that event that just happened. And I want to show you one example of Mashable and that example shows you the search interest after the massive earthquake in Nepal in April and how people search for the two words helping Nepal. And we helped Mashable in that way to visualize that data but Mashable on their side of course gave the context and gave the analysis and gave the background information for readers where they can help victims of the Nepal earthquake and how they can do that. So the important thing here is that Google data is interesting from a point of view that you can visualize it very easily but journalists, academics have to put it in context and analyze it and give a reader the perspective on what that beautiful animated visualization actually means. So I think at that stage where journalists and academics use Google data or data in general the most important thing is to really put it in context. Another example I wanted to show and it is visualization and it is interactive or it was interactive but we had some technical pickups here. So it's not moving at all but you can definitely look it up afterwards. But what we did there, we analyzed a year, a data set of over a year searches about climate change, so air pollution, global warming and other issues and ranked that based on three different approaches. So the first approach was we looked at the top questions being asked in the top 20 major cities. Second approach was we were looking at search interest in each city. So in Berlin, in Paris and Rome, what were people looking for on Google if there would look for global warming, air pollution, etc. And then we were ranking the cities by interest and so we got a really immersive picture of what were people thinking and looking for on Google if they were related to climate change issues and I would love to show you the moving of but it doesn't work. So Google trends data and data in general can't only tell you changes in behavior or changes in interest but it also can tell you how on which level curiosity is actually based. So this is again a very simple example and it's been done by the telegraph actually, looking at in which language do you actually Google. And they analyzed 135 languages in nine cities from 2004 to 2015 and you can see per city the change of the languages, which language was mostly used when people were searching on Google and a really simple idea behind that and so Google trends can show you how the world's most used languages actually change over time. Another example, which I found interesting, looking at Japanese as a language and you can see how the visualization is moving and New York still being on rank two in that case but it's changing over time. So that's the interesting part there that Google data can show you a wide range of data pieces and data sets but if you put it in context you actually see the changes of behavior or interest. And of course elections all around the world are from a data editor point of your data journalist or academic point of view, really interesting because you can analyze what people look for and search for. They search for candidates, they search for topics, they search for parties and they sometimes or this data set sometimes even gives you a trend line on what they will vote in the end. And I hope this visualization works but yes, okay, it works. So we looked at the Republican candidates and looked at the way it changes or the perception of the candidates changes. And I don't know if you see it, so red is Donald Trump and if you look now at another and the second visualization you will see that it's all red. So you see that there's much more interest in Donald Trump than there was before. Another really beautiful visualization, it's like a horse race but it's actually the GOP debate right after, well during the GOP debate. And this is also what Google data can provide. It's this real-time data and the interest and shows the interest in change, shows the change in interest in those candidates. And I think the way it is visualized is actually quite compelling and easy to understand even for a non-data expert or academic. So the question is always how do you get these data sets? One way and that's also the work of the Google Trends team. One way is to go to our open GitHub page and we provide journalists, academics and users with specific data sets that we give them access to. Another way is for example Google public data explorer. We can search for specific data sets, download them and maybe combine them with maps and create really beautiful visualizations. So in order to finish here, our goal and our mission is to collaborate with journalists, to collaborate with academics and to make sure that the data that Google provides from a journalistic point of view is being put in context, is being analyzed not by us but by journalists and academics. And I want to thank you for listening. And if you want to learn more about the News Lab, of course go to g.co.com or follow me on Twitter and approach me later. Thank you.