Clashes of political activism and modern day jokes, “memes”, find a common home in today’s most popular teenage media: Twitter, a social service that allows people to express their thoughts in 140 characters or less. Although Twitter has an inherently introspective focus, there are many times when knowing the mass opinion on such a media giant can be beneficial to students and researchers alike. Chirp is the combination of finding a solution to this problem as well as displaying this solution in a media form that is heavily associated with the teenage demographic.
How it works
Chirp uses sentiment analysis to turn tweets into musical melodies. First, the user will enter a keyword to find which tweets he or she would like to find results for. Once the keyword is entered, Chirp implements a natural language processing algorithm called AFINN; this algorithm takes one tweet at a time, assigning each word in the tweet a positive or negative score based on surveys determining its connotation and public opinion. Once a score has been assigned to each word, the tweet is given an overall score, usually on a scale from negative to positive 13. In more laymen’s terms, the algorithm judges how happy or sad the post will make a normal person feel. The process of determining this integer score for a social media post is known as sentiment analysis. Reiterating this process each time new tweets are posted from around the world creates a huge data set that can be extremely helpful in determining whether a subject is seen as positive or negative on social media. The unique part of Chirp is that it not only collects scores from a sentiment analysis, but assigns musical notes to each score. An interval of scores is assigned to each note on the typical musical staff; as tweet data is sent to Twitter.com, Chirp gets a sentiment score and plays a sound based on this feeling. Higher scores are higher notes, and vice versa. Popular notes show a symphony of social activity and conversation. Searching for a controversial topic such as “Donald Trump” conducts an orchestra of politics and news articles while simple, sweet keywords like “fruit” can bring up a general tone of neutrality and slight happiness.
How many challenges did we run into?
What we loved about making Chirp
The design and code collaboration in this project was ridiculous. There were so many components of the back-end that needed to connect with front-end images and buttons to create such an interactive and appealing display. We loved the way the app's final design and functionality turned out.
The future of Chirp?
In the future, we hope to deploy Chirp and continue to analyze the real-world social effects that are contained within each song Chirp generates.