Do you clean the text (data) before using TextBlob or do you use TextBlob first and then clean the text (from punctuation, stopwords etc)?
1 Answers
I tried both. I seemed to get better results after cleaning, but this was on a small sample size. Try TextBlob(data) and TextBlob(cleaned_data). It helped give a lower score to an article about a car crash. If you know your uncleaned data pretty well you can assess which works better. data.lower() might be good too. That makes it all lowercase. Idk if that helps with TextBlob, but it's what I normally do when cleaning text.
Related