The rates of depression are increasing year by year, having more effect in the young people as never before. We wanted to create a solution that helps people with disabilities giving the proper tools to psychologist to understand the behavior and progress of their patient on a daily basis. Talking with contacts that practices psychology we received an invaluable feedback that motivated and inspired us to make this happen.
What it does
It gives support to psychologists, not only to diagnose but also to prevent, marking a great impact in improving the mood of the patient. It collects data from the patient smartphone with his/her previous authorization, in order to rate their behavior and display that information both to the patient itself and the psychologist. In case of a extreme crisis, it will send an email in order to shorten the time of reaction. The patient can stop the data collection and any time but the psychologist will receive an email. This information can be useful for the sessions, to analyze what happened, when and how.
How I built it
We have developed a mobile application (with Android Studio) for the patient that records all searches and messages, send them to a server hosted in Microsoft Azure, analyzes this information (using a model trained with python) classifying it according to its depressive tendencies and a webpage for the psychologist to recover and analyze the data. For this, we have used the sentiment140 dataset, that provides 1.6 million tagged tweets with the emotional background of it. This dataset is vectorized using word2vec which makes possible converting words into vectors so we can compute the semantic similarity of different words. Then these vectors are used to train a Random Forest Machine Learning model which is finally able to difference between positive and negative phrases. In turn, the psychologist receives through a web application the evolution of each patient, if someone if in a critical condition, the psychologist will receive an alert (developed with Azure Web App and Azure database)
Challenges I ran into
We knew from the beginning that we wanted to help people with depression by analyzing their behavior and giving a evaluation of their progress. But the biggest challenge was to figure out how to make them feel comfortable sharing a information so sensitive and private. Also, how to use it for making their life betters, and after a lot of brainstorming and iterations we find out that the best way of helping them, was helping the people that are making their life better.
Accomplishments that I'm proud of
We honestly think that this can help people with a disability like depression. We're also proud that we didn't die developing it.
What I learned
We have learned that not all disabilities are understood despite the evolution of the world of medicine. We learnt to work hand to hand, integrating many languages and supporting all team members, and we have also found interesting all the things that can be developed with Azure.
What's next for Mokka
Distribute our application to psychologists to help know and treat their patients, giving a more objective point of view and deeper. Later we also want to provide it in hospitals, thus reducing the number of suicides.