Progress: An AI-powered Mental Health App

Abstract

This paper explains the design decisions, development process, and evaluation of Progress, an AI-powered mental health app that tracks continuous patient data. Progress aims to improve patient outcomes by using sentiment analysis and machine learning anomaly detection algorithms to identify valuable trends in patients’ mental health data. The mobile app offers a way for patients to record their daily thoughts in a journal and track metrics specific to their mental health issue. The journal entry text is then run through sentiment analysis to provide a breakdown of emotions, which provides a set of quantitative data that can be evaluated and tracked over time. Lastly, our app uses an anomaly detection algorithm to detect abnormalities in sentiments. When observed over a longer time period, these anomalies can form patterns and help mental health patients realize potential triggers or recurring events that may be linked to volatile surges in emotion. In this paper, we will evaluate the success of Progress -- specifically, the effectiveness of the sentiment analysis and anomaly detection algorithms when applied to mental health data. With 2017 being a prime time for disruption in the mental health industry, Progress has the potential to pioneer a shift to data-driven mental health practices.