Inspiration
What it does
This is a health insurance predictor app that mainly aims at especially helping students at handling their health insurance decisions. We know how taxing it can be for students with no information like us from other countries to be suddenly thrown amongst the wild with zero information with a fast pace. This app mainly focuses on dealing with the problem of lack of information for students. This app mainly focuses on three things.
- To provide you with efficient visualization of how much, people had to pay extra from their pockets due to health insurance they used rather than facing you with lots of readings or contracts. We aim to make it simple and hassle free for you to make a decision and be tension free
- It has a machine learning model which has learned from past data of how much a visit to the doctor due to various reasons, the place, the intensity and many other variables and let's you know a probablistic figure of how much amount of money would your insurance or any insurance is gonna pay on behalf of you and how much it would claim from you.
How we built it
We used Taipy Library to use python to develop the web app which has 4 major functionalities. Initially we used Auth0 as the authentication library to login to the web application. Later, we used Taipy to build the whole web application. We have included two machine learning algorithms and trained these two models with ample amount of data which had brought out results that are very feasble to look at with real world sense. One model is specialized at making near accurate predictions with 4 - 5% error-rate, at calculating the amount of extra money it would cost you, given your scenario of being offered a surgery, or specialized treatment or a prolonged rehabiliation etc. There are many times where we are waiting on the information from insurance agencies to confirm that they are gonna take care of it. But at the end of the day, we will be getting extra bill right when we least expect it. So , this model helps you with the gray zone and reduces your tension. We used Random Forest regression to train the model on a data that had 8 variable dependencies to traain and predict the range that one needs to be ready for in case he has faced with a heavy scenario
Challenges we ran into
Not having much proper documentation or resources to have to look up on. We didn't find much resources to learn more about Taipy and Auth0 as they are relatively new and very few people worked on it. Learning all the nooks the old classic hard way, did make us go back to our roots and dusted up our brains in the modern age. It was a fun experience we had , to rack our brains to learn new interesting stuff
Accomplishments that we're proud of
To successfully integrate the idea of visualizing the model data that satisfies our intention. It solves a problem that we have faced ourselves in the past 6 months. Working together, brainstorming to come up with a solution for the real world problems we faced, made us feel hyped up for the future and also regarding the potential libraries like Taipy hold in this age of big data and utilization of data
What we learned
What's next for Insurance Predictor
We want to continue working on it to increase the no of functionalities it can provide to solve as many problems as it can to all the new age students. We are planning on developing a model that would take in the past data and calculate the future prediction on how much the user would spend if he's gonna continue the same lifestyle else where or with another insurance. We also want to update the maps interface that would help us in specifying us the nearest hospitals on the map at any time, whenever we are in need, that would best help us with the insurance we have . It would show us the hospitals live on map that would cover everything of our insurance for us, or would cost a little bit but give us the best experience we desire for
Log in or sign up for Devpost to join the conversation.