Know Your Audience: Exploring Livestream Chat Through Machine Learning with Charlie Ringer

Video Game livestreaming is an extremely exciting and popular platform for game developers to show off their games and for viewers to consume gamecentric content. Streaming is big business and is only growing more popular, some examples of just how big streaming has become include:

  • Twitch.tv as a whole having more average viewers than all but the largest US cable TV channels[3].
  • Some new AAA titles are looking to influential streamers when executing marketing strategies, in some cases paying huge amounts for brand visibility on stream[1].
  • Livestreaming services such as Twitch, YouTubeGaming, and Mixer being the prime media platform for esports, a $1 billion dollar industry[2].

It is surprising then that, despite its burgeoning popularity, livestreaming has been subject to very little academic study. Even more so, the viewer chat has barely been studied using Machine Learning/Natural Language Processing techniques, despite the huge data available and apparent appeal of understanding the viewership of a stream [4]. In part, this is because the language that is used in viewer chats is very different from properly formatted English and introduces a whole vocabulary of symbols which have a complicated and rich meaning, meaning that traditional natural language processing tools are either inapplicable or require heavy modification.

This talk, therefore, will introduce livestream viewer chats and the unique challenges that go along with studying them as well as presenting early finding on advancing our understanding viewer chats through Natural Language Processing techniques. In doing so, I attempt to make as few assumptions about the meaning, especially sentiment, behind stream specific words and instead 1 learn a model of the meaning of words through co-location with other words and inference from outside sources.

[1] Ea reportedly paid ninja $1 million to stream ’apex legends’. https://www.engadget.com/2019/03/13/ea-paid-ninja-1-million-to-stream-apex-legends/. Accessed: 2019-07-05.
[2] Newzoo: Global esports economy will top $1 billion for the first time in 2019. https://newzoo.com/insights/articles/newzoo-global-esports-economy-will-top-1-billion-for-the-first-time-in-2019/. Accessed: 2019-07-05.
[3] Why study video game livestreams? https://www.charlieringer.com/blogs/Why-Study-Streams.html. Accessed: 2019-07-05.
[4] Mehdi Kaytoue, Arlei Silva, and Lo ̈ıc Cerf. Watch me playing, i am a professional: a first study on video game live streaming. Proceedings of the 21st international conference companion on World Wide Web, pages 1181–1188, 2012.

Return to Schedule

Sign Up For Free Tickets

Grab your tickets from eventbrite


Send us a message below