As you can see in the screenshot, I can count the values of each possible value for the sentiment column, and then I can group the same data by the brand and the sentiment, but the values are obviously incorrect, since they don't add up with the total values for each sentiment.
brand_sentiment_train.sentiment.value_counts()pivot_df = (brand_sentiment_train.groupby(['brand', 'sentiment']).size() .unstack(fill_value=0) )pivot_dfI'm not smart enough to figure out what the issue is that is causing the groupby method to incorrectly group the sentiments with the brand, so I am hoping someone else can see what the error is. All I know is that it seems to be an issue with the groupby method.
I tried to group the data by sentiment and brand, and I was expecting to have the total amount of each sentiment add up to the same amount of each sentiment that I determined for the sentiment feature individually, but instead, the values are way off, and it appears that when I try to group by both sentiment and brand that somehow the sentiment is not being counted correctly.