CareCost

CareCost Logo
CareCost Input Screen
CareCost Graph Screen
CareCost Welcome Screen
CareCost BMI Calculator
CareCost Information Screen

Inspiration

We were inspired to create CareCost after finding a dataset on Kaggle that tracked people’s healthcare costs billed by their insurance provider based on a variety of conditions. Based on this data, we created an easy-to-use app that lets users input certain metrics to get a personalized prediction of their future medical expenses. We hope that by giving users a convenient way to understand their upcoming healthcare costs, they will be encouraged to make responsible fiscal decisions and plan ahead.

How we built it & What it does

To build the app, we first developed a prediction model in Python. Utilizing scikit-learn, a machine learning library in Python, we calculated a regression model for the dataset. The model had an r-squared value of 0.75, which is relatively strong for a financial model. The weights that were outputted by this model are utilized in our iOS application.

Using XCode and Swift, we developed an app that lets users input their age, BMI (we implemented a BMI calculator for convenience), state of residence, smoking status, sex(at birth), and number of children. The app inputs these metrics into the prediction model through the weights that were calculated from earlier, and displays a chart with a slider so users can see their estimated annual out-of-pocket healthcare costs in the next 1 to 10 years. After giving users an idea of how much money they will need in the next X years, they will be taken to a screen that provides a variety of advice on how to save money for healthcare. Other credible sources are also linked for the user’s benefit.

To make the UI more appealing, we used Coolors.co to narrow down a visually pleasing color template. We also designed a logo for our app (a heart balloon with a dollar sign) and animated it throughout the screens to make the app more engaging.

Challenges we ran into

A challenge we ran into the beginning was creating an accurate regression model of the data. We experimented with our code and it took much tinkering and research to create a model that had any predictive power. Without an effective model, there was no way we could make an app.

Accomplishments that we're proud of

We are very proud that we have created an app that has the potential to create real change for people. By making an app that is accessible and easy to use, we hope more people can be aware of their finances regarding medical expenses. Medical expenses probably are the most important costs in life and being aware of how to save money for this is crucial. When people can not pay the out-of-pocket costs they can become drowned in debt and will be discouraged to seek medical help in the future, putting their own health at risk.

What we learned

We learned that simple datasets can be transformed into powerful tools of good. Raw rows, columns, and numbers in a CSV file can be transformed into empowering information that can make an impact.