MKTR.AI Co-Founders Building Twitch Automation Tool


  • Share on Google+

Let me start by saying

We weren’t trying to hack Twitch or scrape their valuable data. We’ve recently become intrigued by Twitch for it’s potential for businesses and entrepreneurs, but more on that later.

Let’s talk about the Twitch automation tool

This small project is due to my curiosity. I want to know what types of accounts people follow, and whether I can discern any insights from their available information.

Do people who watch Ninja play Fortnite, watch pro athletes play NBA2k19, and follow Gary Vee happen to all fit in a specific segment in terms of other preferences and behaviors? If so, what preferences and behaviors do they share? (This is the basis of customer segmentation)

And…

Most importantly, who are the people most likely to benefit AND be interested in our content, live streams, and produced videos?

We want to grow our following from zero to thousands. Who are the people we need to connect with to make that happen?

We don’t get to those answers in these videos but build the baseline of an automation tool.

What do we have so far?

So far, we have written a script to pull the profile links of everyone who follows a specific profile. Once we get the code buttoned-up, we’ll share it here and on GitHub.

First Twitch Live Stream Session

In the first video, we figure out the basics of the Twitch follower page HTML and figure out how to automate the scrolling for an infinite scroll page as well as pulling all of the links from the page. This was a rough coding session.

It had been a while since I used Python + Selenium + Chromedriver, and it showed. There was a lot of trial and error.

I started in a Jupyter notebook for testing and early iteration and closed the session by transferring the code to Microsoft Visual Studio Code and formatting the code in functions.

Second Twitch Live Stream Session

In this video, Brandon, our CTO, cleans up my sloppy code and then runs into some issues with Selenium. He spent his time improving the precision of the script. His mission was to:

  1. Only capture the URLs of the profile followers
  2. Automate the scrolling of the page, which proved difficult with the infinite scroll page, within a div.

What now?

There’s still a lot to do on this project, but we plan to continue to chip away at it.

This post will be updated as we move further along with this Twitch Automation tool.

If you’d like to see us work on something in particular, please comment below or reach out to us on Twitch.

Thanks for reading and watching! 

  • How to Protect Your Content & Turn Thieves Into Partners | Mike Nemke
    Reply
    Author
    How to Protect Your Content & Turn Thieves Into Partners | Mike Nemke

    […] If browser automation or scraping interests you, check out my Co-Founder & me struggle-bussing through building a Twitch automation tool […]