1 in 5 seekers of mental health services in Philadelphia do not receive the care they are seeking (slightly worse than the national average). Our app aims to increase accessibility of mental health services to increase the positive outcomes associated with this community. We wanted to explore the application of AI in assessing and determining the level of care needed by a person in crisis.

What it does

Our app receives information from user input and displays appropriate resources. The first assessment determines if the user is looking for services for themselves or for someone else. After this determination, users are given choices to help them determine the type of services they are seeking. Some of the outcomes include: connections to hotlines, outpatient providers, and inpatient facilities. Each of the providers in our data set have been evaluated to meet certain criteria. They must be board certified as well as accepting new patients. Our data set was built from information gathered from the Philadelphia Division of Community Behavioral Health's directory of providers. Users also have the option to submit a vocal recording describing their emotional state. From the recording, a sentiment level is determined by AI which assesses mood instability. This sentiment level is then used to assess the level of care required by an individual in crisis.

How we built it

Python and Beautiful soup were used to aggregate data related to our project. Our application was developed using the LAMP architecture. JavaScript was used to interface between our database and and display query results in a table. We used bootstrap to implement design elements. For the Artificial Intelligence component of our app, we used voice recognition software to transcribe the audio recording submitted by the user. Our application is accessible to users of all languages. Our app uses Google Translate API to translate foreign speech.

Challenges we ran into

Integrating the applications Staying awake and alert Centering elements with CSS

Accomplishments that we're proud of

Incorporating AI into our application with Google Cloud APIs Building a dynamic website Pair programming with GitHub Bringing awareness to mental health

What we learned

A functioning web app incorporating AI can be developed in 36 hours Some members of our team learned Beautiful Soup Others acquired new skills building dynamic web applications with JavaScript We gained a functioning knowledge of Google Cloud APIs including Google Cloud Natural Language API which was used to determine sentiment analysis, Google Cloud Translation API which was used in the language accessibility feature of our application, Google Cloud Speech-to-Text API used for the transcription, and Google Cloud Compute Engine used as the web server for our application

What's next for Mind Hub

Increasing the amount of providers in our dataset Refining the application of AI Presenting more crisis intervention resources Creating a location based system that displays a map of providers Implementing a search function

Share this project: