VMS 198 – Experimental Interface Design
Below I outline my reflection of one of my most important classes thus far in my Duke career — Experimental Interface Design. This class allowed me the opportunity to create my own working prototype of my concept business — Critic. Below is also an artifact of my critic presentation that I made at the end of this course along with a brief description of what the project is about.
Critic is a text summarization web application for multimedia reviews. Critic is an original product prototype that uses Natural Language Processing (NLP) and Sentiment Analysis on film, tv, games, and music text reviews across the internet to give the user more insight on the content they consume in a more concise simpler way.
Critic is for the consumer who doesn’t want to read full articles of reviews but also doesn’t just want a blanket numbered score. With Critic the user can both see the score that reviewers rated certain content but also see, at a glance, why. Critic uses adjective noun pairs to give the user a condensed summary of what the review is about based upon nouns and the most used terms used in all reviews in general. Examples may include, story… fascinating, characters… heart-warming, dialogue… unconvincing, production… weak, etc.
Not only can you get a highly summarized yet contextual overview of content before you consume it, users also have an option to get personalized recommendations based on what THEY deem important and valuable, not what critics. If you complete the attribute profile, rating how much you care about certain attributes of content, then Critic will be able to recommend things to you based upon your profile. For example, if you heavily care about action in film but don’t really care about dialogue and character development that much then based on the sentiment analysis of reviews Critic will recommend certain content that aligns with your own values.
*Critic at this point is a prototype. All of the companies mentioned are not involved with this project.