Understanding COVID-19 Outcomes by County: An End-to-End User Journey in the Cloud

 

Workshop Leaders

Eric Tassone, Associate Professor at UVA School of Data Science


Casey Lichtendahl, Associate Professor at UVA Darden School of Business

 

Time: 11:15AM- 12:45 EST

 

In this workshop you will participate in an investigation that attempts to explain and understand COVID-19 outcomes by county. We will marshal common theories about what drives outcomes, identify and prepare relevant data, and analyze the data with the aim of explaining what drives outcomes.

Our data will come from publicly available sources, and we will conduct our investigation as an end-to-end journey on an integrated, cloud-based platform. The investigation will rely on geospatial data and be conducted at the county level from a public health perspective. Along the way we will look for disparities in both outcomes and potential drivers, such as economic inclusion, comorbidities, and vaccination rates. In our analysis we will model relationships between drivers and COVID-19 outcomes while explaining and quantifying the contributions of the drivers.

The platform for our investigation will accommodate ingestion, cleaning, reshaping, joining, exploring, and visualizing the data we select. It will also enable powerful analytics and modeling of our data in a “low code” (i.e., SQL-like) environment. We will deploy state-of-the-art Machine Learning methods and, importantly, Explainable Artificial Intelligence (“XAI”) techniques to communicate our findings. The platform is simple to spin up, easily accessed, and fosters a team-based workflow -- an important attribute because data science is a team sport.

*While not necessary, some familiarity with SQL would be a plus for folks who want to code along with us. In that spirit, familiarity with parts of Google Cloud Platform (e.g., Google Cloud Storage and BigQuery/BigQuery Sandbox) would also be a plus for folks wishing to code along, although it is not required.*

 

There is no need to register for this individual session, to register for Datapalooza go here.