We are a passionate group of data scientists and we identified a need for an application to improve ICU bed allocations during covid.

What it does

The system predicts the current covid active cases in the healthcare system based on the previous week's data, then computes an elective surgery schedule for a particular state that maximizes doctor preferences given the prediction of the number of ICU beds for COVID patients. Provides a state level and hospital ICU bed schedule tool combining physician and policymaker input.

How we built it

We built the predictions data set using pandas and then built a multi-layer perceptron in PyTorch to do the covid-19 case predictions. We then built an allocator script to do the specific state bed allocation

Challenges we ran into

Manipulating the data set was more challenging than expected to build the training set for the model

Accomplishments that we're proud of

The model for covid cases was much more effective than anticipated! The Predictive Bed Allocation Algorithm improves on the bed allocation implemented during covid, by*reducing delayed elective surgery* and improving the balance of ICU bed allocation between covid patients and elective surgeries.

What we learned

Data preprocessing is difficult but rewarding!

What's next for Predictive Bed Allocation Algorithm

Continued development and possible academic paper submission!

Built With

Share this project: