We read a paper about disability-adjusted life years (DALY) which is a measurement of burden of disease in a region, and that inspired us to relate this to the severeness of COVID-19 in each region of the United States and build models to predict that.
What it does
To create a predictive pandemic model to risk stratify populations and generate actionable dashboards to guide resource allocation and planning.
How we built it
We leveraged the following data; Updated COVID19 County-level Cases and Deaths, Population Demographics (Size, Age Stratification, Ethnic Makeup, Density), Disease Burden (Disease Adjusted Life Year Scores, Medicare Data) We investigated population trends reported in the news regarding COVID 19. Also, We applied various statistical and machine learning approaches to publicly available data, including Multivariate Regressions, Random Forest, Lasso Regressions, Neural Nets. We deliberated on the possible use cases for these efforts, Disease Knowledge, Risk Stratification, Resource Allocation, Pandemic Tracking and so on.
Challenges we ran into
The data has a lot of mistakes, nulls and formatting errors that we have to clean before the computer can use it. We also need to meet online every few days in order to make sure we are making progress.
Accomplishments that we're proud of
We accomplished the project in an extremely short time with 6 team members. The coordination of our team is fantastic because we are all passionate about this hackathon.
What we learned
We showed the statistically strong effects of public disease burden data from Medicare and Disease Adjusted Life Year Scores on COVID 19 cases, deaths and hospitalizations. We also created map data to point out the most vulnerable regions of the United States for the virus.
What's next for Daly Tally
To create a modular platform for pandemic tracking, planning and resource allocation by leveraging our data and models. We plan on creating county level compartmentalized simulations that use our data approach to create visualizations of population interactions, population risk composition, and pandemic growth metrics.