Inspiration

In emergency situations, immediate first aid knowledge is needed. Google searches and youtube videos can be cluttered with ads and irrelevant information. To solve this, we create our PatchPal web app companion to offer users immediate first-aid advice.

What it does

On our web app's landing page, we provide a search bar where users can ask first-aid-related questions, such as "How do I perform CPR?" or "What do I do if I got a cut?" Our trained model will provide instructions on any issue the user has. For more common procedures or questions, our web app has pages dedicated to readily-made information on "How to prepare your first aid kit," "Performing CPR," and "Emergency U.S. Hotlines." These provide instructions and resources, and the CPR page even offers the user recommended songs for CPR.

How we built it

We used OpenAi for the chatbot model and fine-tuned the responses with several first-aid question kaggle datasets. We used Flask to host our web app, with Python as our language. We also used Spotify API to connect user data for recommended CPR BPM songs.

Challenges we ran into

Our main challenge was finding appropriate datasets to fine tune our ChatGPT model.

Accomplishments that we're proud of

We're proud of trying out many different APIs. We are also proud of fine-tuning the model and finding datasets for it.

What we learned

We learned how to use multiple APIs in tandem and fine-tune models.

What's next for Patch Pal

We want to refine the responses, by making them better formatted, and generally improve the user interface. We hope to add more pages with easy access to common first-aid problems.

Built With

Share this project:

Updates