2 min read
Back in August of this year, Google launched a set of prediction models for the ongoing COVID-19 pandemic in collaboration with the Harvard Global Health Institute. Known as the COVID-19 Public Forecasts, the models provide fairly accurate projections of COVID-19 cases, deaths, ICU utilization, ventilator availability, and other important metrics for US counties and states. Now, the Forecasts have been updated by both organizations to expand its projections beyond the US and into the rest of the world.
Trained on volumes of data from Johns Hopkins University, Descartes Labs, the United States Census Bureau, and a host of other sources, the updated COVID-19 Forecasts will now serve as a critical resource for first-responders in healthcare, the public sector, and other relevant organizations.
The Forecasts work by allowing organizations to conduct targeted testing of affected regions and intervene in a timely manner. The pandemic is a rapidly-evolving monstrosity, after all, and the use of data to predict its intensity in certain regions and determine how to prepare for it is a very helpful asset to have indeed.
For instance, healthcare workers could use these predictions to help them in resource planning for equipment like PPE, and activities like staffing and scheduling, Moreover, health departments could use these forecasts to determine testing strategies and identify areas at risk of an outbreak.
Google is looking forward to provide support to countries beyond the U.S. as well, hence the recent updates. It is now introducing the COVID-19 Public Forecasts for Japan, which is based on publicly available data like the COVID-19 Situation Report in Japan. The company is also making the forecast models customizable, so that they can be extended to other problems and datasets.
More than 100 Alphabet employees contributed to the development of the COVID-19 Public Forecasts, fine-tuning and tweaking it on a daily basis to make it the comprehensive model that it is today.
