Having told you a little bit about our moral dictionary, the reason why we're building this dictionary is because we want till we get moral frames at scale. We don't just want to look at moral frames in sample. No, we want to analyze moral trajectories in global on lines that are happening at 15 minute intervals around the world. Okay. Thank you, Freddie. That was great. So why does this all matter? Let me explain to you this. Events happen everyday. We know that news organizations report above those events. We also know there's a theory behind it that of course events lead to reporting about these events. But we also know that reporting about events can actually trigger subsequent events. So these are dynamic involved between how news organizations, journalists, report about events, and what events subsequently may have the busted probability of events that may happen part of, unrelated to reporting, but part we need to report. So if you have information on global level, global scale, how events unfold. As you can see in this map, that's a map of a time where and when events happen. If we have information on how reporting about these events are framed, morally framed, include moral information of which we know, there are motivation relevant and that we drive those events. As you can in this map, that's a map that shows you how these small frames unfold over time. We can apply some statistical advanced magic called spatial temporal modeling and extract relationships between these moral information, it can be small frames, and events that happen. Then we can predict reporting of these events that should work because people report about events. But the interesting question is, is it possible that we actually use the reporting and then predict the probability of subsequent events? Think about protests, demonstrations. Usually people have a reason why they go the street and protest about a certain issue that they care about. I would even argue most of this issue can be morally framed. As I said earlier, if you see that someone is unfairly treated and this unfair treatment is unjustified, that makes you upset. That might be the reason why you actually go to the streets and protest. So if we have this information about how information happens or framed morally, we may be able to use this information and then actually predict when, where, and in what intensity those protests may occur. Let's move this to a different domain. As Freddie has explained, our Moral Narrative Analyzer, MoNA, is applied in the news narrative domain but we can also apply MoNA of that in the fictional narrative domain such as movies that we all like that much. Think about that at a regular normal standard day, all the Hollywood studios, they keep about 30-100 film scripts everyday. People from all over the world send them film scripts. That means, a text file of hundreds of pages with texts that tells people who says what with what intention in what setting. That's how they should do it. Because they get hundreds of films of everyday, it is a problem how do you select film scripts that maybe good scripts that lead to well performing, interesting movies in which one might be the little bit artsy still great movies but may not be the next big blockbuster movie. As you can imagine, this has directly financial implications. So how it's done now largely, not exclusively, but largely, those film scripts are sent to film students, they read it. Those are different film students, they change from day to day sometimes. They get paid but they read these scripts and make a gut decision whether this might be an interesting movie or not. Then there's a new selection going on, and those scripts that are selected move on to the next round where another group of experts will talk about these films. So MoNA comes in here that if we know that moral conflict is the interesting part of the movie, the crucial scenes in a movie all are about moral conflict, not in all movies but in most of the movies. If we have a tool that can help us computationally process a large amount of text data such as film scripts and helps us to understand what are the character relationships, how other character relationships framed by moral information, how does this moral information translate into a moral conflict between two or groups of characters, how does this dynamic of moral conflict unfolds over the course of a movie, then even more important if you can track and store all this information. It's not lost from one student to the next student that will look at this whole scripts. So we can actually learn from these information in the past and then maybe able to make even predictions, how these movies perform economically, that's really interesting.