Optimood : Transforming Health Data into Practical Advice and Insights

In light of the significantly increasing rates of depression, an application was designed that automatically derives mood insights from personal health data. Patterns were found by analysing correlation coefficients and extracting feature importances from a random forest regressor. The effectiveness of the application was examined in a small-scale usability test, where participants were asked to register numerous mood-related variables. In general, participants were able to effortlessly interpret and identify themselves with the generated insights. The extent to which the insights delivered any new or valuable information was less than expected, however. Future research should examine how more advanced and deeper patterns can be discovered and visualised accordingly.