The only way to understand this data is to work backwards.
The detections show a series of words in tweets that were flagged as possible political misinfo.
All of these words were part of a larger NLP vocabulary for Twitters Machine Learning model for political misinfo🧵… twitter.com/i/web/status/1…
Read the article before asking any questions. The article explains the larger context.
rubymediagroup.com/twitter-artifi…
The theme that emerges from this data is that political misinformation only existed on the right. I see no evidence that political misinformation existed on the left. Several data scientists reviewed the data and confirmed this assessment to be true.
Another issue is the partisan media category in the detection list. I see no evidence that any left leaning media outlets were included in this list. All partisan media outlets are from conservative media platforms.
Keep in mind I am only reviewing one small sample of data in one category.
What about Civic? Health? Abuse? I don’t have that data. What do those n grams look like? What words are included on those lists? To understand the full picture- that information is necessary.
Share this Scrolly Tale with your friends.
A Scrolly Tale is a new way to read Twitter threads with a more visually immersive experience.
Discover more beautiful Scrolly Tales like this.