Applying one simple regex to the text body of the previous one million tweets creates the most amazing result.
re.findall(r'((?:[A-Z][a-z]+\s?)+)(?<!\s)', full_text) turns into a list of what is currently trending on Twitter. Simple, powerful and elegant.
Conversation
Replying to
I don't understand how a pattern matcher can be linked to finding "Trending", which is purely regression analysis (?). Unless of-course all the valid strings in your data are just already ranked trending topics.
1
Replying to
You are correct. I should have clarified a bit. Basically it is extracting proper nouns that are capitalized so it is really picking up people and places that are currently ranked high in activity.