Using Natural Language Analysis to Assess the Sentiment of Web Based Comments

 

Between 2015 and 2018 I periodically assisted with the natural language analysis component of the Rap Almanac, using IBM Watson natural language tools.

This project encouraged me to develop a prototype for a tool for assessing the sentiment of comments for the sites I manage. I would very much like for my sites to have a comments section, but incessant spam makes comments sections expensive to operate.

That being said, the IBM APIs are straightforward and fast, so implementation of a comments sentiment assessment tool can be achieved from scratch, with one or two week’s work by a medium level programmer. I’d be interested in assisting with such a project should someone have the spam problem under control.