9th Session- Integrative Lab: NLP

In this part of our Integrative Lab we will select two of the key videos identified with help of our SNA, and analyze the sentiment and emotions contained in the comment sections of the videos. We use NLP from IBM Watson for this.

1) Select two (2) nodes from your previous network that were shared by both channels. Look up their URL links in the Gephi database (for guidance see our previous tutorial here: UCCSS_LAB_SNA: Network Measures (24min) ). (if, and only if, there are less than two shared nodes, feel free to choose any two nodes, but show and explain this)

2) Take each of the two URLs (one after the other) and collect all comments of each video. You can do this the webscraper.io tool or a simple spreadsheet (here is a tutorial for how to do this:  UCCSS_Lab_ScrapingComments-1.pdf Download UCCSS_Lab_ScrapingComments-1.pdf), or use a public online app (like  http://ytcomments.klostermann.ca or https://seobots.io/bots/youtube-comment-scraper  but, as always when you rely on code of others, this might or might not work...).

(if, and only if, you do not get a result from this scraping, feel free to choose other nodes, but show and explain this)

3) Obtain the csv files with the comments:

4) Copy the >commentText< from the first 500 comments. Do a semantic analysis for >sentiment< and >emotions< using IBM Watson Natural Language Processing: https://natural-language-understanding-demo.ng.bluemix.net 

5) Analyze the >sentiment< and >emotions< results for the comments of BOTH of your selected videos. What are some of the conclusions you can draw? Do they both contain similar sentiments and emotions? Looking at the videos and quickly browsing the comments, what do you think of this NLP result? Discuss.