Tags
1. Introduction
Nowadays, data science is becoming more and more important in the Entertainment Industry. Different data science methods are used in this fields to improve user experience. So what is the trend and future for data science in this industry? This is the topic we want to talk about in this post. We will take game industry as a case analysis for detailed explanation.
2. Game Industry – A relatively new but popular sub-industry for entertainment fields
Video games and online games dominate the game industry. And almost every big game company will have data research teams to support their operation. Data positions in game company can be classified into three main types: data analysts (MBA students dominating), data engineers or database administrators (CS student dominating) and data scientists (Phd student dominating). In the future, the most changes will be taking places for all of the data guys.
2.1 24 hours data need 48 hours to proceed – Big Data Curse
First of all, as more and more people are involved in games, the data volumes are increasing significantly with unbelievable speed. Before game companies mostly use excel, traditional database and easy data presentation tool for their work. But as time going on, data guys in game company finds out that “Oops, our one day’s user data has to take us 48 hours to proceed, we need revolution!” So they decided to take new tools for their data. This is the real story for Riot games.
Riot Games, Inc. was founded as an independent game studio in 2006 by Brandon “Ryze” Beck, and Marc “Tryndamere” Merrill in Los Angeles.The company announced its first game, League of Legends: Clash of Fates, in October 2008, and released the game in October of 2009 as simply League of Legends. [1].
Riot Games Logo [2]
Before the game is growing so fast, they simply use traditional and simple data tools for data work. However, as more and more players are fascinated with this game, their data became really big data. There is a picture to compare Riot games data situation between the past and now. In this picture we can see that the number of tables before is 180 but now is 1200! The increase is more than a double. Also there is no pipeline event per day, but now the number of pipelines is 7+ billion. The tools and environments are also changed significantly. So as the data trend going on, in the future, their data will become huge and new technologies and tools must be applied. [3]
Riot Games Data Increasing Comparison [4]
Data is infinite, but human efforts is limited. How to solve this conflict? That’s a question for the future.
2.2 Data is more informational and helpful than you know
Data in game industry now is used most in a statistical way. For example, as we talked before, “Flappy Birds” is taking your location data as into their analysis process. They want to know the geographical distribution for the players of their game. But in the future, more values inside data will be exploited for better use.
Let’s go on with Riot Games to be our case study topic. Now, Riot is using data to improve player’s experience, because they take players’ feelings as the priority. But do you know data can have infinite information for us to explore?
First of all, Riot wants to take “log chatting” data into consideration as their next step. Players tend to chat a lot during games, and their chatting records can reveal a lot of useful information. For example, let’s imagine that you are playing a champion that you think it is the worst. So during the game, you are complaining about that champion all the time with your teammates. And they will also reply you with their attitudes about the champion. So after collecting your guys’ chatting data, sentimental analysis can be done to help Riot improve their champions. Also, in your chatting records, some basic information of yourself will be revealed too. That helps the game company to grasp more of their user characteristics. Interesting, right?
Second, detailed game play even tracking will be done more carefully. Riot games will track your actions and reactions in the game to know more about their players’ behaviors. This makes them easier to catch your feeling and do prediction on you. Definitely, this can improve your experience in game and of course, this can also help them earn more profits.
Data science provides a win-win situation in entertainment industry!
2.3 Data Center Everywhere
Let’s imagine that all the things we mentioned above have been already perfectly completed by Riot, the players are still complaining. Why? It is data latency!
Huge data centers for Riot need to be built everywhere in the world, at least in four to five main countries’ core cities. Especially, for Asia countries, they have limited Internet access but high game demands. Physical solutions must be established together with our data plans mentioned above. So data centers become the biggest support.
In the future, more and more data centers will be built by entertainment industry to satisfy huge data need all over the world and reduce data latency.
3. Conclusion
In this passage, we use case study for Riot Games to explain future trend for data science in entertainment industry. More changes about data will take place and be seen by the world in the future. Data scientists, are you ready?
[1] “Riot Games.” Wikipedia. http://en.wikipedia.org/wiki/Riot_Games
[2] “Riot Games.” Wikipedia. http://en.wikipedia.org/wiki/Riot_Games
[3] “Oozie @ Riot Games.” SlideShare.http://www.slideshare.net/mattgoeke/oozie-riot-games
[4] “Oozie @ Riot Games.” SlideShare.http://www.slideshare.net/mattgoeke/oozie-riot-games