For this hackathon, we wanted to build something that could have a positive impact on its users. We've all been to university ourselves, and we understood the toll, stress took on our minds. Demand for mental health services among youth across universities has increased dramatically in recent years. A Ryerson study of 15 universities across Canada show that all but one university increased their budget for mental health services. The average increase has been 35 per cent. A major survey of over 25,000 Ontario university students done by the American college health association found that there was a 50% increase in anxiety, a 47% increase in depression, and an 86 percent increase in substance abuse since 2009.
This can be attributed to the increasingly competitive job market that doesn’t guarantee you a job if you have a degree, increasing student debt and housing costs, and a weakening Canadian middle class and economy. It can also be contributed to social media, where youth are becoming increasingly digitally connected to environments like instagram. People on instagram only share the best, the funniest, and most charming aspects of their lives, while leaving the boring beige stuff like the daily grind out of it. This indirectly perpetuates the false narrative that everything you experience in life should be easy, when in fact, life has its ups and downs.
What it does
One good way of dealing with overwhelming emotion is to express yourself. Journaling is an often overlooked but very helpful tool because it can help you manage your anxiety by helping you prioritize your problems, fears, and concerns. It can also help you recognize those triggers and learn better ways to control them. This brings us to our application, which firstly lets users privately journal online. We implemented the IBM watson API to automatically analyze the journal entries. Users can receive automated tonal and personality data which can depict if they’re feeling depressed or anxious. It is also key to note that medical practitioners only have access to the results, and not the journal entries themselves. This is powerful because it takes away a common anxiety felt by patients, who are reluctant to take the first step in healing themselves because they may not feel comfortable sharing personal and intimate details up front.
MyndJournal allows users to log on to our site and express themselves freely, exactly like they were writing a journal. The difference being, every entry in a persons journal is sent to IBM Watson's natural language processing tone analyzing API's, which generates a data driven image of the persons mindset. The results of the API are then rendered into a chart to be displayed to medical practitioners. This way, all the users personal details/secrets remain completely confidential and can provide enough data to counsellors to allow them to take action if needed.
How we built it
On back end, all user information is stored in a PostgreSQL users table. Additionally all journal entry information is stored in a results table. This aggregate data can later be used to detect trends in university lifecycles. An EJS viewing template engine is used to render the front end.
After user authentication, a journal entry prompt when submitted is sent to the back end to be fed asynchronously into all IBM water language processing API's. The results from which are then stored in the results table, with associated with a user_id, (one to many relationship). Data is pulled from the database to be serialized and displayed intuitively on the front end. All data is persisted.
Challenges we ran into
Rendering the data into a chart that was both visually appealing and provided clear insights. Storing all API results in the database and creating join tables to pull data out.
Accomplishments that we're proud of
Building a entire web application within 24 hours. Data is persisted in the database!
What we learned
IBM Watson API's ChartJS Difference between the full tech stack and how everything works together
What's next for MyndJournal
A key feature we wanted to add for the web app for it to automatically book appointments with appropriate medical practitioners (like nutritionists or therapists) if the tonal and personality results returned negative. This would streamline the appointment making process and make it easier for people to have access and gain referrals. Another feature we would have liked to add was for universities to be able to access information into what courses or programs are causing the most problems for the most students so that policymakers, counsellors, and people in authoritative positions could make proper decisions and allocate resources accordingly.