The visualization doesn't interest me nearly as much as how Twitter classifies something as "toxic". It seems subjective enough that one person's "toxic conversation" is another person's "sick burn". Unfortunately, the post doesn't contain any explanation or links regarding their classification algorithm(s).
So toxic, but what defines toxic, how were they categorised etc? Just can't take the term seriously anymore after all the buzzfeed videos educating me on American politics that's been shoved into my feeds.
It seems unfortunate that the goal was to build a good internal marketing video, rather than a useful user interface that could be used for navigating Twitter?
The final viz looks awful. In the attempt to make it look like a tree, you did get a tree but lost the informational value. I thought the earlier iterations had less noise and more signal.<p>There are plenty of books on visualization theory. Exposes pitfalls.
This is an awesome related API by Google: <a href="https://perspectiveapi.com/" rel="nofollow">https://perspectiveapi.com/</a><p>Also check out this Kaggle competition: <a href="https://www.kaggle.com/c/jigsaw-toxic-comment-classification-challenge" rel="nofollow">https://www.kaggle.com/c/jigsaw-toxic-comment-classification...</a>
It’s amazing to me that people look at an online community such as twitter and then determine if it’s toxic or not based on observed behaviors. They’re just people. People are toxic on twitter. Twitter is not toxic. Twitter is exactly the way that people who use it want it to be.<p>How do you expect to change twitter without changing people? It’s an exercise in futility. You can regulate it for eternity and it will still reflect the people who use and engage in the site.