Inspiration
We are high school students and know how stressful life can be. Sometimes we just need a place to rant without feeling burdened, judged, and express our emotions. We need something to help us through the challenges of life, so we created this mental health buddy to reduce stigma and help others receive mental assistance with affordability and accessibility. Cheer up!
What it does
First, it has user authentication and requires sign up and sign in. Then, it has stripe api for payments which are affordable, only 2 dollars. Then, users can go right to the chat and start getting assistance and Therapy. From the messages, the machine will send over human like advice, encourage the client with memes, quotes, and music as well with natural language processing.
How we built it
We built TherapYEET starting with an ML model, trained off the IMDb dataset . The model was built with scikit-learn. Secondly, we developed a web-based API for user authentication and activation of account based on payment. After that was done, we built an Android app that integrated with the web API. The Android app featured Stripe for payments, although we used a test API key so no one would lose money during the test. We used Azure for the cloud.
Challenges we ran into
We ran into challenges choosing an appropriate dataset for the task, because the IMDb dataset does not reflect humans engaging in day-to-day conversations. While we found another usable dataset more corresponding to what we needed, we were unable to use it due to time restrictions and lack of information on how to read the dataset quickly. Also, with the lack of wifi, it was hard to keep editing and working throughout.
Accomplishments that we're proud of
We are very proud that we were able to get through the challenges with Android Studio and implement the Stripe API, SQL authentication, and nice features. We're also proud that although our ML model is not perfect due to the lack of perfect datasets and lack of time, it is sufficient for our purpose and can help mental health of others. We are also proud that we implemented the in the cloud machine learning with Azure and were able to put the pieces together.
What we learned
We learned that you have to be mindful of the type of text in a text dataset if you are making an application that uses ML and is focused around text. We also learned how to quickly deploy Flask applications on Microsoft Azure. Finally, we learned how to write a chat application using ListViews in Android, how to integrate Stripe with Android, and how to integrate a web API with Android.
What's next for TherapYEET
TherapYEET will gain true credit card transactions for low-cost service, by using a real Stripe API key instead of a sample. Hardcoded values will be cleaned up into centralized files. Next, we will use the dataset we found in order to have better accuracy with the application, and find more appropriate memes and songs. We will finally perfect the audio playing situation in the app. We will then branch out and move to more platforms, making a web UI and an iOS apps and improve the UI.
Log in or sign up for Devpost to join the conversation.