Inspiration

Whether it is your own family or a stranger, seeing someone in a hospital for days is indeed heart-wrenching. We are inspired to use the dataset at hand to understand what factors are essentially connected to a person's length of stay in hospitals. Doing so would allow doctors and medical staff to leverage these factors to improve patient and clinical experiences.

What it does

The app provides a visualization by race, gender, and age group to understand hospital encounters. Then, there is a model that anyone can use to enter a person's race, gender, age, medications, etc. to predict how long they might have to stay in hospitals. An easy-to-use app provides interactivity, user-friendliness, and convenience for users to generate insightful information from data.

How we built it

We built the visualization dashboard with Tableau. Then, we used R libraries to understand potential predictors of length of stay in hospitals. After doing so, we saw a correlation between the time spent in the hospital and most of these factors, so we fit a model (backward-selection model) using R code to predict the length of stay in hospitals. Finally, we used the shiny package in R to create an entire app that allows users to make real-time predictions with a clean interface, and flexible input options with combinations of dropdowns, numeric fields, and sliders.

Challenges we ran into

The major challenge was that this was both of our first hackathons and we hadn't worked on such a comprehensive and long project before, so we were somewhat unsure of the techniques and tools we had to build. Furthermore, neither of us had worked with Tableau or R-Shiny, so it took good research to come up with these ideas and execute them effectively.

Accomplishments that we're proud of

The dashboard and the app serve as unique ways to interact with data in a user-friendly way. Also, one of us had a more statistical background whereas the other had a more programming background, so we are proud to have complemented one another throughout the hackathon.

What we learned

Learned to produce solutions under time constraints, to work independently, to conduct independent research, and to relate our research to real-life problems.

What's next for Health and Nostalgia: Predicting the time spent in hospital

Given a more comprehensive dataset, perhaps we can further enhance the predictive power of our model!

Built With

Share this project:

Updates