What mental health needs is more sunlight, more candor, and more unashamed conversation. You don’t have to struggle in silence and medication is not always needed. Unhealthy mental health contributes to an unhealthy lifestyle, which in the short run causes people to be alienated from their peers because of perceived unattractive personality traits or behaviors and in the long run drives a person to commit suicide. According to the National Institute for Mental Health, over 90 percent of suicides have depression or another mental disorder as factors and knowing friends and family members with similar conditions, we are tempted to provide them with a more innovative, portable, accurate, efficient and cheaper way of therapy.

Cognitive-behavioral Therapy (CBT) is proven to be as effective as medication in the treatment of depression and anxiety, or in developing self-awareness, in challenging underlying assumptions, and in viewing the world from a more holistic perspective. With an ever-evolving cognitive conceptualization, CBT treatment plans target several issues to establish attainable goals, facing fears and avoiding them, and establishing challenging yet reasonable goals. As a process of self-evaluation for therapy, it is important to describe, accept, and understand rather than judge themselves or others and with several CBT methodologies, the most efficient sessions can be achieved.

One such reflective methodology is using thought journals, cognitive restructuring or reframing, and guided discovery. Implementing these methods allows us to thoughtfully consider the world and our own ideas and experiences holistically by understanding our strengths and weaknesses in order to support our learning and personal development. Through such self-awareness, we can assess our growth and effectiveness and change course when necessary, thus leading to a healthier lifestyle.

What it does

TherapAI is an application to recognize one’s distortions in thinking that are creating problems, and then to reevaluate them in light of reality. Through our highly interactive way of journaling using different kinds of input such as speech, visual art, text, multi-select, and radio options, we allow users to experience a real therapy session whenever they want to, without having to wait for months, unlike how therapy sessions work here. Our interactive and reciprocally active responses are generated from complex speech recognition, real-time video recognition, sentiment analysis in text recognition, and their answers in several scientifically proven tests such as Rorschach Inkblot Test and color tests that use Carl Jung’s archetypal theory for colors. The data once collected from the self-reflective journal is then cleaned, analyzed, and used for prediction to tailor future questions more appropriately so the user ends up feeling grateful, compassionate, balanced through their thorough and interactive self-reflection.

Through sophisticated cognitive and behavioral therapy, TherapAI reduces depression by giving users tools to challenge the negative thoughts and override them with more realistic and positive thought processes. The solution focuses on changing patterns of thinking and beliefs that are associated with, and trigger, anxiety. This application is a guided, interactive app that takes in user input in 3 phases and allows users to self-reflect on themselves, and creates individualized personal reports with graphs and charts to quantify their emotions. Along with quantitative data, we have qualitative data included in the report so that both the users and therapists are able to understand their health.

TherapAI is a useful tool to address the emotional challenges as it can help you manage your symptoms of mental illness, prevent a relapse of mental illness symptoms, treat a mental illness when medications aren't a good option, learn techniques for coping with stressful life situations Identify ways to manage emotions, resolve relationship conflicts and learn better ways to communicate, cope with grief or loss, and overcome emotional trauma. Backed up with research (also found in the “Cognitive Behavioral Therapy” portion in our website, such a method helps combat depression, anxiety disorders, phobias, PTSD, sleep disorders, eating disorders, obsessive-compulsive disorder (OCD), substance use disorders, bipolar disorders, schizophrenia, and many more conditions.

How we built it

1) The frontend was created using React Bootstrap along with HTML, CSS, JS. 2) The authentication was done using Firebase 3) The backend was made using Django, using the Models.db database. 4) The connection between React Bootstrap frontend and Django backend was achieved using REST APIs. 5) Once the user is logged in, through a variety of interactive inputs, we used the following tools for the analysis of their answers: 6) Voice recognition: 7) Real-time facial recognition: 8) NLP sentiment analysis in text: 9) Using the input from their voice, video, and text, we detect the mood/emotion of the user and modify our questions and generate report analysis 10) The final pdf was created using the PyFPDF framework in python with a combination of graphs from matplotlib and seaborn.

Challenges we ran into

1) One of the challenges we ran into was connecting the frontend with the backend: it was challenging to collect user input from different kinds of input such as text, sliders, multi-select, and voice and use them all to efficiently analyze the data. To resolve this issue, we created a REST API that transmits information from the frontend to the backend and vice-versa. 2) After collecting the data and generating graphs, we also had a tough time generating a PDF document so that the data collected is easily accessible. To overcome this issue, we combined the graphical images generated using Matplotlib and Seaborn with the PyFPDF framework.

Accomplishments that we're proud of

We’re proud of being able to do the following: 1) Successfully being able to implement voice and real-time facial recognition in our website while maximizing user experience. 2) Developed a REST API to connect the frontend and backend. 3) Being able to use matplotlib and seaborn to generate graphs based on real-time data.

What we learned

1) Collecting data from facial, voice, and text to perform sentiment analysis for mental health purposes and generating visually appealing graphs using matplotlib and seaborn. 2) Being able to develop REST APIs tailoring our personal needs. 3) Being able to automatically generate PDF summarizing and analyzing the user’s data. 4) A lot of machine learning and data visualisation!!!!!

What's next for TherapAI

1) Making our phases more sophisticated by adding more nodes in our decision tree so it is more specific to a wide range of cases. 2) Making our inputs more interactive by reducing the number of text inputs to allow our users to have more of a “real-life” experience. 3) Making the charts in the PDF more specific, and being able to give the users to customize the items they want to see in the report i.e., they can modify their timelines as it currently shows it for their entire history with the application.

Built With

Share this project: