Sentiment analysis of the biggest YouTube feud of all time
What does the data say?
Cover image by Wikimedia Commons: Image 1 and Image 2
James Charles vs Tati Westbrook
Let me introduce you to one of the biggest YouTube controversies ever - James Charles vs Tati Westbrook.
James Charles is a beauty guru, YouTuber, and makeup artist. He has close to 26 million followers on YouTube, and his videos have been watched over 3 billion times.
The influencer has found himself in the centre of multiple controversies in the past.
Two years back, James was involved in one of the largest scandals in his life when fellow beauty YouTuber Tati Westbrook uploaded a video about him.
In the video, she calls him out for being manipulative and using her, and accuses him of predatory behaviour.
Popular beauty influencers like Jeffree Star and Shane Dawson have been vocal about their support for Tati after she released her video about James.
This isn't the first time James has been involved in a controversy like this. Every time a new scandal comes to light, he finds himself under fire and is heavily criticized.
However, he also has a loyal fanbase of teenagers who support him despite any allegations thrown his way.
I used to watch a lot of his makeup tutorials when I was younger, and always found the James Charles situation interesting.
So like any other data scientist, I decided to do an analysis around his feud with Tati.
I wanted to collect enough data to understand what people really thought about him, and whether public opinion on James has changed over time.
You can find all the codes for this analysis here.
First, I analyzed Tweets to look at overall sentiment around James Charles for the past two years:
The red line indicates negative sentiment, and the green line indicates positive sentiment.
For such a popular influencer, James has a lot of negative sentiment surrounding him.
From the visualization above, we can see a huge spike in the number of Tweets with the hashtag #jamescharles around May to June.
Normal days seem to have only around 100 Tweets, but this time period has over 2000.
Let's take a look at sentiment breakdown around this time period:
May - June 2019
There is a spike in both negative and positive sentiment on May 11 2019. We see 10X more tweets than we normally would.
What happened on May 11 2019?
After doing a quick Google search, I realized that May 11th was the date when beauty influencer Tati Westbrook uploaded a video painting James in a negative light:
Tati's video went viral, and millions of people all over the world (even people who didn't know who James Charles was) were suddenly aware of the drama.
In Tati's video, she accuses James of being a bad friend. Tati sells vitamins and she was upset that James, her close friend, was promoting competitor vitamin brands.
She mentions that she had promoted him multiple times and helped with his rise to fame, but he never did the same. According to Tati, James was a liar and a master manipulator.
She mentioned that James would talk to her about coercing straight men into going out with him. According to Tati, this was predatory behaviour, since he was abusing his money and fame to get younger guys to go out with him.
After watching Tati's video, the Internet was filled with many different opinions.
People were divided, and weren't sure who to believe.
On one side, there were diehard James Charles fans who refused to believe he would do any wrong. They wanted to know why Tati, a 37 year old woman, was making a video bashing a 19 year old boy.
On the other hand, there were people who already disliked James for his previous scandals and genuinely believed Tati when she called him a manipulator.
Overall sentiment surrounding this situation varied a lot, so lets build a word cloud to better understand it.
It appears as though our initial assumption was right.
People were pretty divided over the entire situation. On the left, we see words like love, support, and believe. This indicates that a bunch of people believe James Charles and will still watch his videos, continuing to love and support him.
On the right (negative sentiments), we see words like manipulate, lied, stupid, and men. From here, we can infer that some people think he is manipulative and a liar for hurting Tati and preying on straight men.
Now, lets take a look at the previous chart again:
We can observe another spike in both positive and negative sentiment on May 19 2019.
What could have happened during this time? It looks like all the drama was eventually dying down, and people seemed to be forgetting about the entire incident until May 19.
A quick Google search led me to this:
James Charles uploaded a 41 minute long video called No More Lies on YouTube defending himself against Tati's allegations on May 19 2019, which is why we suddenly see a spike in Tweets and mentions during this time.
If you look at the chart above, you'll notice that we see a spike in both positive and negative sentiment during this time.
However, we don't know what people are saying.
Are most people on James' side after watching his new video? Or do they still think he's a liar?
To understand this further, I decided to take a look at the top comments on James' No More Lies YouTube video.
YouTube Sentiment Analysis
We don't know if a majority of people are siding with James or Tati.
The negative comments could be against either one of them.
Lets take a look at the top comments of the No More Lies YouTube video to understand the situation better:
Most people seem to be reacting positively to James' Bye Sister video. They mostly seem to be siding with James now, and negative comments are targeted towards Tati.
Next, I build word-clouds to further analyze the YouTube comments:
The word-clouds seem to confirm this belief. There are now far fewer negative comments aimed towards James Charles, though many people still believe that he should be arrested for his predatory behaviour.
The positive comments indicate that a lot of people believe James over Tati, and are surprised that a 37 year old woman dragged a 19 year old boy for not promoting her brand.
James Charles has been involved in many other controversies since then, and the most recent ones have caused him to lose partnerships and brand affiliations.
I won't be analyzing the more recent scandals since there is just too much ground to cover.
I also want to clarify that I believe both parties could have handled this situation better, and I am not taking any sides with this article. All the analysis performed here is purely for entertainment purposes.
I performed this analysis because I was looking for a fun project to do, and I wanted to practice my Python and sentiment analysis skills.
I used to watch YouTube videos a lot when I was younger, and kept up with all the controversies surrounding YouTubers.
I thought it would be interesting to do an analysis around a topic I already was interested in. Plus, it really did help hone my skills in data analysis and visualization.