SDS 244: Data Science in Entertainment

Welcome back to the SuperDataScience Podcast ladies and gentlemen, super excited to have you back here on the show. And today we continue our series of episodes about data science in various industries to help you excel in those industries or inspire you to come up with even cooler, even better ideas for your industry.

And so far in this series we’ve done, the following industries, we’ve done data science in healthcare, retail, mining, education and banking. And so there’s five episodes in total and those are episodes number 216, 220, 228, 234 and 238 in case you’ve missed them, you can find them on the SuperDataScience show. And today in this episode we’re talking about data science in entertainment. Very exciting episode ahead. Some cool examples coming up. So let’s dive straight into it.

The entertainment industry entails commercially popular activities for people to amuse themselves with, particularly cinema, television, radio, theater, music and gaming. So we’re all familiar with the entertainment industry. In fact, as a species, we’re moving more and more towards a world all full of entertainment. According to some estimates, the revenue of the entertainment industry in 2018 was $2 trillion, 2 trillion with a T. If you recall from our previous conversations, for example, that’s higher than the mining industry combined, the total of the mining industry and the entertainment industry comprises about 2.5% of the world’s GDP. So a massive industry, lots of people are employed here. Let’s have a look at how data science and artificial intelligence are applied in the entertainment industry.

Application number one, targeted advertising. With the world of entertainment expanding exponentially, there’s more and more types of entertainment and kinds of content. And it’s important to make sure that whatever you’re advertising as a business, as a company or as a data scientist, when you’re helping a business, whatever you’re advertising to your customers, it has to be relevant because there’s so many different options. Some might not fit certain people and certain people might be looking for specific things. And that’s why you want to make it your advertising as specific as possible. In that case, is going to be a win win both for the business and because therefore you spend less on advertising or use ineffective advertising and also a win for the customer because you enhance the customer experience and therefore they’ll come back to. And the example here is YouTube. So I’ve got a very cool example for my personal life and it’s to do with my mom. So my mom uses YouTube on her iPad and she is very protective about other people using her iPad specifically on to do, to search things on YouTube to watch videos.

And the reason for that, when once we discussed, the reason for that is really cool because my mom watches, you know, maybe she’ll watch some cooking videos or some self-development videos, or some other things that might be of interest to her on YouTube. And then she will wait for YouTube to give her recommendations. So she actually has a really cool relationship with YouTube where she watches videos and then, by putting likes and thumbs up, thumbs down and subscribing to channels and selecting the videos and the content that she watches, she helps YouTube understand what her preferences are, what her goals are, what she wants, what kind of person she is, and then, she works together with YouTube to iron out that perception that it has of her. And then she consumes the content that YouTube recommend her.

She’s very excited about the new things that YouTube suggests and recommends to her. And so whenever somebody else like my dad picks up the iPad and searches something for himself in her account in YouTube, that completely disrupts her whole relationship with YouTube because all of a sudden it starts recommending completely different content that is not, it’s not what she wants to know, what she’s after. So she’s very protective of her iPad, of their relationship. I think, I really enjoying seeing how old generations, regardless of age, are embracing this new world we live in. Where artificial intelligence can actually help us in our search and consumption of entertainment and actually make our lives more interesting and fun and easier because we don’t have to do that search ourselves. So there’s a short example from my personal life and now we’re moving on to application number two of data science in entertainment.

And the application is creating specific content. So the world’s actually moved well beyond just using data science and AI to recommend existing content to people. Companies are now actually creating content based on artificial intelligence insights and data set. And here’s an example. Here are some mind blowing facts about Netflix. This is really crazy stuff. So in 2018 Netflix spent 13 billion dollars between 12 and 13 billion dollars on content and how crazy, that is just like mind blowing. $13 billion on content. And 85% of this is actually for new original series and movies. So to give you a quick comparison, Netflix subscribers in 2018 got approximately 82 new feature films to watch, whereas Warner Brothers created only about 23 in new films, and Disney only about 10. So those numbers are quite approximately because these were predictions made earlier in 2018 in the middle of 2018, but nevertheless, compared that 80 new films by Netflix versus about 23-25 by Warner Brothers and 10 by Disney.

In addition to that, in terms of television series, Netflix produced or acquired an incredible 700 new or exclusively licensed TV shows, including more than a hundred scripted dramas and comedies and dozens of documentaries and children’s shows. How crazy is that? We’ll link to all these stats in the show notes where you can read more about this, but the numbers speak for themselves. $13 billion spent on content in 2018 by Netflix. And what they’re doing is they’re actually taking all of the data from hundreds of millions of users that they have and identify what is exactly of interest, what are people searching for, what kind of shows do they like, what kind of actors do they like, what kind of directors do they like? What combinations of those things do they like? And based on that, they know, they’re very well informed what kind of content to produce. They have the budget, they put in 13 million, 13 billion, excuse me, $13 billion. And off they go. They started creating videos that are laser specific to what people actually want.
So that’s creating specific content. 

Number three application of data science and AI, number three, and this is gaming AI. So Ai has been very closely linked to gaming, to all sorts of games starting from Deep Blue beating Garry Kasparov in chess, in the game of chess in 1997 to AlphaGo winning the world championship of Go of the game of Go, which is much more complex than chess, exponentially more complex, in 2017 a decade before it was actually predicted to happen. And now our artificial intelligence beating humans in the StarCraft II, a very popular game in the world right now. This was, happened in early 2019. So the reason why artificial intelligence is very closely linked to gaming in that sense, because gaming provides an enclosed environment where we can see how artificial intelligence can cope with a limited set of rules and perform, in a very… Beat humans at something that they’re good at.

And not only does that show us that artificial intelligence is advancing and that we can apply same types of artificial intelligence in business, to solve enclosed environments and complex problems, but also changes how we actually perceive games. But that’s not the only way AI is applied in gaming. Ai is actually applied on the other side as well to compete with humans to make the challenge more interesting, to make the challenge more exciting, to speak of entertainment, to have a lot of people to be entertained, not just go through the game too easily or not be able to pass the game because it’s too complex, but actually have a worthy opponent. And so an example here is a game called The Last of Us, which it was released in 2013 and it sported some of the best gaming artificial intelligence to date. The game received wider claim from players and critics and sold over 17 million copies making it one of the best-selling video games of all time.

And a lot of that success was attributed to or is attributed to the artificial intelligence that was present in the game. And that was back in 2013. Imagine how far we’ve advanced now in 2019.

Okay. Application number four of data science and AI in entertainment and this is trend prediction. So very similar to how Netflix creates content and chooses which videos to create and where to invest its budget into that can be done across many areas of entertainment to actually see what people are interested in. And that’s a very different approach to what companies used to use back in the day where it’s gut field, where when they’re thinking or you know executives or managers or whoever’s putting the strategy together have a gut feel of what is going to be next, what’s going to be the next big hit. Now we can use data science for that.

Example here is Shazam. We all know Shazam is the APP that you can download on your phone and you click a button and it will tell you what song is playing in the room you’re in or wherever else. Like the song you’re hearing. You can just through a tiny little brief sample, it will tell you exactly the name of the song so you can save it to your Spotify or wherever else you are listening to music and listen to it later. Well that’s not the only thing that Shazam does. Shazam actually also uses the data of what people, what song is that people shazaming and then it uses those insights to predict what kind of genres are going to be popular next. What’s going to be the next big hit. What kind of music are people actually interested and are going to be interested in the future. And, no wonder that Shazam was acquired by Apple in 2018 for $400 million.

So they’re definitely doing something right and also puts into perspective that our world of entertainment isn’t as simple as it seems. In the background there’s all these AI and data science algorithms working to help enhance our experience and actually guide not only individuals in their direction of what content they’re going to be consuming, not only companies in what, in terms of what content they’re gonna be creating, but our society, our world as a whole in terms of where we’re moving, what direction we’re moving in terms of what’s going to be the next big hit, what’s going to be popular, what genres we’re going to be listening to and things like that. So quite a very deep philosophical thing, if you stop to think about it for a second.

And moving on to application number five of artificial intelligence. This is a pretty cool one, Visual effects rendering. So last year there was a really cool movie. I haven’t seen it yet, but after reading this, I’m really looking forward to watching it. It’s called Avengers: Infinity War. If you’ve seen it, then you probably remember the character called Thanos. I hope I’m pronouncing that well. And so what happened there? Right. First of all, Thanos is if you haven’t seen movie is like this big purple dude and he actually it’s… Basically is bigger than a human and his face is bigger than a human. So this special effects that are involved in order to create this character. So yes, there’s an actor walking around, but then there’s a lot of special effects overlaid on top in order to create this physiology and his appearance. So basically what happened this time is that the visual effects team used an AI algorithm trained on high resolutions scans of the actor’s face in order to bring the character of Thanos to life in Avengers: Infinity War.

The AI was so good, it could detect individual wrinkles and automatically map the resulting face renders onto the character’s body. And what this did is it allowed the actor to work alongside the other stars rather than being in front of a green screen alone. And also it allowed the director to see the results in real time. So critical acclaim for the movie was very high, much of it focused on how believable Thanos was and it grossed over $2 billion making it the fourth highest grossing movie of all time. So they go a quite an unexpected application of artificial intelligence and those are only just five examples or five areas where data science or an artificial intelligence can assist in the entertainment industry. There are plenty and plenty more. If you’re inspired by these, I highly recommend for you to check out some more and read further into it.

It’s a huge field, it’s a huge industry and there’s lots of fun, exciting things going on in the space of data science and AI.

And on that note, thank you so much for being here today and I hope you enjoyed this little course into the world of entertainment. If you’re interested in getting the show notes for this episode, then head on over to www.www.superdatascience.com/244. That’s www.superdatascience.com/244. And there you will find more links where you can read more about these examples and also any other material mentioned in the podcast. And on that note, we’re going to wrap up. I look forward to seeing you back here next time and until then, happy analyzing.