Rackspace Technology, UC Irvine, Scientific Systems and Plan4Co Debut COVID-19 Forecasting Model

Rackspace Technology, an end-to-end multicloud technology solutions company, has partnered with University of California, Irvine, Scientific Systems Company, Inc., and Plan4Co to work on a project that underscores the effectiveness of Amazon Forecast in creating COVID-19 forecasting models with greater accuracy.

“This partnership constitutes the kind of collaborative enterprise necessary to advance our forecasting capabilities in a more fine-tuned and nuanced way that may well yield important actionable policy initiatives,” says David A. Snow, Ph.D., distinguished professor emeritus of the UC Irvine School of Social Sciences. Dr. Snow has led team-based homelessness research throughout the U.S. and the world, including Paris, Sao Paulo, Tokyo, and Orange County, CA.

Leveraging the serverless data processing solution built on Amazon Web Services (AWS) by Onica, a Rackspace Technology company, the team—comprised of experts in medical, academic, data science and forecasting fields—demonstrated that hospitalizations and deaths can be more accurately forecasted by utilizing the Rackspace Technology solution compared to other commonly used forecast models. The effort showed an improvement in the short-term accuracy of the forecasts by leveraging timeseries data ingested into a data lake built on AWS and pipelined into an advanced machine learning algorithm.

“We’re proud to play a role in helping humanity overcome the worst public health challenge we’ve seen in the modern era,” says Tolga Tarhan, CTO, Rackspace Technology. “We’re humbled that our data scientists were able to put our extensive AWS and AI/ML experience to work to predict future needs and strengthen planning efforts—both of which have never been more critical.”

The project is composed of two major innovations. The first innovation, known as COVID DSAFE, is intended to deliver improved data synthesis and advanced forecasting techniques. Powered by Amazon Forecast—a fully managed service that uses machine learning to deliver highly accurate forecasts—COVID DSAFE produces operational forecasts weekly at the state, county and zip code level.

The second innovation, known as COVID PrivyTRAC, is a contact tracing system that runs on AWS.  This system provides for COVID contact tracing with advanced privacy and security features, addressing common privacy concerns via advanced statistical spatiotemporal models.

Sources: Press materials received from the company and additional information gleaned from the company’s website.