Taylor Swift Sentiment Analysis

Conclusion

To frame this conversation, I will be referring to changes in overall sentiment with respect to the Positive Sentiment curve.

In the period before the Eras Tour announcement, the Positive sentiment % is typically above 50% which indicated that the overall sentiment is largely positive.

In the period after the beginning of presale to just before the start of tour, the Positive sentiment % is still largely above 50% but dips below that during certain periods of time indicating the sentiment became more negative and/or neutral. Looking at the negative and neutral curves, we can visually confirm this is the case.

In the period after the opening shows, the Positive sentiment % is around 50% which is lower than during the first period analyzed.

Looking at the Sentiment % Overtime (All Tweets) graph, it's easier to see the Positive sentiment changes overtime indicating how sentiment became more negative and/or neutral overtime. The sentiment spikes don't map directly to the dates listed in my hypothesis so there is something else going on.

I noticed that there were several unrelated tweets in the dataset that were included because they mentioned her account so they were likely skewing my results. Better screening of the Tweets and excluding irrelevant ones (i.e. Tweets related to the opioid crisis in Toronto and Tweets about the Ukraine crisis that mention Taylor's account should have been excluded) would have likely made the sentiment spikes correlate more to the timeframes I had listed.