Over hot pot, we found a dataset on Kaggle that predicted no shows at doctor's appointments. From subsequent research, we discovered that 20-30% of patients miss their appointments, contributing to clinical inefficiency. We wanted to expand this, and build a tool that helps solve this problem. We implemented an algorithm we found in a paper, and integrated it into a website.
What it does
Schedules appointments in an optimized fashion.
How we built it
We wrote the website with Flask and Bootstrap. The backend was created using Python, and a basic scikit-learn linear regression was used.
Challenges we ran into
Learning new technologies, finding a project that was doable under the time constraint and had enough data.
Accomplishments that we're proud of
Modifying and implementing the algorithm. Scalable framework. Finally figured out forms on Flask.
What we learned
Don't reinvent the wheel if it still works.
What's next for Patient Scoot
Google Calendar integration? Improved UI and error handling. Improving ML model with more data.