MoodMerlin is a web application for sentiment analysis: determining, with permission, how users are feeling based on their posts' tones and patterns. Through the use of an OAuth access token, interested users are prompted to log into and authenticate their Twitter accounts. Users can then run MoodMerlin on their own tweets. MoodMerlin pulls recent tweets from the user's twitter using Twee.py and parses these tweets with the DatumBox machine learning API, which classifies the user's mood in each tweet as 'negative', 'neutral', or 'positive'.
The DatumBox API is accessed through a Python-dependent backend, which is able to be leveraged by Amazon AWS cloud severs. The classification modes are then graphed over time and returned to the user, equipping them with powerful quantitative trends that can be used to inform doctors' treatment plans for anxiety, depression, and other mental health issues if the user so chooses.
What's next for MoodMerlin
In the future, we would like to establish machine learning training sets that would allow us to classify by unique emotion (anger, joy, loneliness, etc) rather than overall mood (negative, neutral, positive). We would also like to launch Android and iOS apps for users' ease of access and mobility of data if users choose to use their data as part of a formal treatment plan with their doctor.
AJ Hinsvark - Freshman - Aerospace Engineering - Embry-Riddle Aeronautical University - First Hackathon
Sean Holden - Senior - Computer Science - Embry-Riddle Aeronautical University - First Hackathon
Jon Rach - Sophomore - Software Engineering - Embry-Riddle Aeronautical University - SwampHacks 2016, UHack 2016
Courtney Thurston - Freshman - Computer Engineering - Embry-Riddle Aeronautical University - SwampHacks 2016, MHacks: Refactor 2016