In this growing world with pandemics and other irregularities, mental health problems are a very common issue and has already taken many lives. Better mental health can not only lead to a person's betterment but also for the society. As an estimated 52.9 million people are currently experiencing mental health issues, it’s obvious that the majority are unable to pay for and access professional help. This is where our website, Mental Health First comes in.

What it does

Our website, Mental Health First is available around the clock to provide short-term aid to those who are unable to immediately receive treatment from health care specialists using features like music recommendations to enhance their mood, reading amazing quotes, and a lot more. We use a variety of proven, noninvasive methods with our features which are implemented after users take a quick diagnostic test using our ML model which predicts their mental health state with their voice tone sample and gives out necessary aids needed. The first step of a person's mental health journey is realizing that they need help. Our website assesses a person's mental state and gives them early steps to improve it.

How we built it

For the front-end of the web application, we used HTML, CSS, and JavaScript to create an appealing user interface that people can use to assess and improve their mental state. Our diagnostic exam was created using python machine learning to analyze a person's voice and predict their mood.

Our model is trained with a MLP Classifier from the sklearn library using data of voice recordings to detect human emotions in audio files. While the training and implementation was successful at first, data corruption of the file lead to an oversized ipynb notebook before we could record the demo.

RIP EmoDetector.ipynb

Challenges we ran into

We were struggling to integrate our machine learning model into the web application as it's our first time making prediction model like this one.

Accomplishments that we're proud of

We were able to make a machine learning model with around 70% accuracy during the duration of the hackathon which can predict a user's mental health state with their voice samples. Also we were able to make a web application where user's can get necessary short term aid to get out of mental health problems like depression or stress.

What we learned

We got to know about machine learning and learned to make models with it. Also got to know about flask framework and learnt to develop with it.

What's next for Mental Health First

Continue development to help those who need it's access with mental health professionals based off of their results. Finalize all features and launch a finished website with updates as needed for our growing user base.

Built With

Share this project: