Inspiration
We believe that burnout can have multiple negative side effects on our lives, especially for healthcare workers. They work around 6 to 12 hours per day with mixed night/day shifts, which might cause a lack of sleep, lack of meals eaten, increased stress/fatigue levels, and less work efficiency. As an employee, less work efficiency means an increase in mistakes and a slower work pace. As a human, lack of sleep/meals and increased stress/fatigue levels can lead to depression, hopelessness, increased rate of heart disease, high blood pressure, and a higher possibility of sudden arrhythmic death.
What it does
We want every healthcare worker to work healthily and efficiently. Our team wants to create a machine-learning model to predict burnout levels for each healthcare worker. Then, we can use it to help them schedule breaks in between shifts and develop healthier lifestyle habits to avoid burnout.
How we built it
We built this project using Python's built-in SKLearn and Pandas libraries and utilized a gradient-boosted regression model to analyze our data which allowed us to predict burnout as well as stress levels, fatigue, and efficiency. We gathered this data by contacting real healthcare workers and having them fill out a Google form. After extrapolating the data using a Java program, we fed it to the machine learning algorithm and used Matplotlib to visualize the results of the learning process.
Challenges we ran into
Some challenges we ran into were collecting useful data and implementing machine learning. Because we had never applied our knowledge of machine learning before, creating a project revolving around ML was a daunting task. First, we had to understand what good data is and how to format questions to create a successful model. Then, we had to learn how to apply that data to predict future results and visualize said discoveries properly.
Accomplishments that we're proud of
We are proud of the data that we got because although our data set was simulated because of the uniqueness of our idea, we used real data from healthcare workers that we extrapolated along with some coding to create a simulated dataset. Learning what machine learning does and how to create a product through it is also something that we are proud to accomplish. In conclusion, the entire project and its ideas were new to all members of the team, so the communal learning and experimentation process is something we are happy to have experienced.
What we learned
We researched the correlations between burnout and different aspects of work and used these correlations to get our own data for our model. We also learned how machine learning works and how to implement proper practices with training large sources of data.
What's next for Healthcare Schedule Helper
We are looking forward to applying this project to hospitals, med-tech companies, and any other company that has problems with burnout. We want more data to feed to our machine-learning model so that it can be more accurate in calculating burnout levels. We can utilize text recognition to realize whether a word is positive or negative, resulting in the user being able to respond with words to express their feelings and allows us to determine their risk level too.
Log in or sign up for Devpost to join the conversation.