YT Live Chat Sentiment Analysis
Chat Extraction
Sending chunks of chat to server
Cleaning of data (emojis and empty spaces and irrelevant links and spam)
Analysis
Send back data to live
Dashboard Elements:
Positive / Negative -> Graph & Meter
Most repeated words
Indicate the streamer to interact with the user (questions in a chunk)
Views
Revenue
**subscribers gained while streaming
Elapsed Time
Algorithm
for indicating the streamer to stop.
Chat rate flow at the last chunk of data sent and question tags of last data.
=> Sentiment Analysis of YT chats : - Sentiment Analyis - Sentiment -> binary classification * Stanford Sentiment Tree Bank * fine grained (50%) 1. Rule based 2. Feature based 3. Embedded Try all of them and see which one is best. (47%) -> Optimise the best approach through small things:
1. emoji consderation 2. ... ...
Images
Emoji Sentiment Map
Reconfigured emoji Dataset into CSV format
Labels for Categorizing
