Sentiment analyser for youtube comments

I am trying to implement a sentiment analyser for youtube comments.

Specifically, I will fetch all top level comments from videos (max 3-4 pages) for a search query. Then I will deduce the sentiment of each comment. Now here I need a bit of help. Currently I did this with the help of open source API (takes sentence and returns sentiment). Can I do better than that?

Secondly, my end goal is to sort all videos based on comments (video with lot of nice comments should be on top). Could anyone please help me here with some algorithm to accomplish that thing.

Try using the VADER sentiment Analysis package in Python.

It is very easy to use and generally works well with social media texts.

It is lexicon based sentiment analysis technique, which not only gives the sentiment of the comment but can also give a score (between -1 and 1) for each comment based on how positive or how negative the comment is. You can later use this score to sort the videos.

© Copyright 2013-2019 Analytics Vidhya