If you’re familiar with Twitch TV, then you should be familiar with their chat. It’s entertaining to say the least, particularly in the very popular channels that have 5-50k viewers. Somebody will say something in chat that others end up copying and pasting over and over. These messages then turn into sort of Twitch “memes” and will spread to other channels too.
From watching this happen, I started getting curious what were the most shared messages on Twitch? Since Twitch uses IRC as their backend for all the stream chat, it’s pretty easy to plug into. I made an IRC bot that will listen to the top 30 channels streaming and gather statistics on messages sent and the words of each message.
After about a week collecting data here is what I found.Total messages processed:12,673,979Unique messages:7,441,845Unique words:2,636,727Peak messages per second:1808Top 50 messages:
|Total messages processed:||12,673,979|
|Peak messages per second:||1808|
|Top 50 messages:
||Top 50 words:
It’s not too surprising that about 40% of messages are duplicate. The chat isn’t exactly known for quality discussion.
The entire dataset, top 1000 list, and the code I used is all open. Feel free to hack/download/check it out.