A Marine Logistics Base May Be the Warehouse of the Future
March 29, 2021
ODU, Hampton Roads Community Foundation preparing residents for hurricanes through virtual event
May 20, 2021A Marine Logistics Base May Be the Warehouse of the Future
March 29, 2021ODU, Hampton Roads Community Foundation preparing residents for hurricanes through virtual event
May 20, 2021VMASC Develops Real-Time Platform That Predicts Spread of COVID-19 in Virginia
Written by ODU PR Staff
April 17, 2020
As COVID-19 approaches its forecast peak in Virginia, researchers at Old Dominion University's Virginia Modeling, Analysis and Simulation Center (VMASC) have developed a real-time platform that predicts the spread of the disease down to the city and county level.
The Virginia County COVID-19 Daily Case Total Forecaster was developed by VMASC researchers Ross Gore and Christopher J. Lynch, to provide a platform where the public can explore COVID-19-related model forecasts.
"One model gives a seven-day forecast that is updated daily," said Gore, an assistant research professor who leads the Data Science & Predictive Analytics lab at VMASC. "In addition, the model provides a forecast of the age range of individuals with COVID-19, and what their case outcomes are expected to be."
The platform also provides a machine learning model which identifies tweets from individuals within each county who are reporting symptoms.
Gore added the platform was designed to be as transparent and accessible as possible. "For each model it includes, it provides an explanation of the methodology it applies that is geared toward a general audience," he said.
The project was taken on independently by VMASC as part of its mission to develop computational models that help solutions in complex areas. However, several local decisionmakers, including Jim Redick, director of Emergency Management for the City of Norfolk, have begun to take interest in the platform.
The model uses an array of existing data sets, including county-level information published and updated daily by the New York Times, data from the American Communities Survey (ACS) and information about hospitalization, intensive care and mortality rate from COVID-19 published by the Centers for Disease Control.
The ACS dataset includes age-range demographics at the county level, which can be correlated with the CDC hospitalizations dataset to estimate an outcome for each case.
"In general, counties where a larger portion of the population are older are at greater risk for more severe outcomes," said Lynch, a senior project scientist who leads the Data Analytics Working Group at VMASC.
In addition, data from Twitter is included in the analysis in affected counties. Posts with tweets that mention "fever" or "cough" are collected daily, then run through a machine-learning algorithm to classify only those tweets of individuals self-reporting symptoms.
The research platform allows anyone to explore, at a very local level, how the number of reported cases is increasing or decreasing in different areas of the state, based on different models. The forecasting models show that some communities, such as Norfolk, have had more success leveling off the reporting of new cases than other localities.
The researchers also hope that the platform can aid other COVID-19 modeling efforts, featuring a section that points to more than 130 publicly accessible external data and modeling resources.
Gore added that this platform is not intended to be the definitive, county-level COVID-19 forecast.
"This is a research platform for exploring models," he said. "We believe the information provided by the platforms can be useful to the public, but our goal is to explore and explain how different models make predictions and forecasts."
VMASC is an applied research center of ODU focusing on innovation, workforce development, and industry ecosystem engagement programs leading to digital transformation. VMASC's research staff of more than 70 research faculty, scientists, support professionals and students perform scientific research, develop computational models and create information-to-insight and digital engineering solutions in areas such as digital shipbuilding, spaceflight and autonomy, and digital health and health equity.
Connections:
VMASC Develops Real-Time Platform That Predicts Spread of COVID-19 in Virginia
Written by ODU PR Staff
April 17, 2020
As COVID-19 approaches its forecast peak in Virginia, researchers at Old Dominion University's Virginia Modeling, Analysis and Simulation Center (VMASC) have developed a real-time platform that predicts the spread of the disease down to the city and county level.
The Virginia County COVID-19 Daily Case Total Forecaster was developed by VMASC researchers Ross Gore and Christopher J. Lynch, to provide a platform where the public can explore COVID-19-related model forecasts.
"One model gives a seven-day forecast that is updated daily," said Gore, an assistant research professor who leads the Data Science & Predictive Analytics lab at VMASC. "In addition, the model provides a forecast of the age range of individuals with COVID-19, and what their case outcomes are expected to be."
The platform also provides a machine learning model which identifies tweets from individuals within each county who are reporting symptoms.
Gore added the platform was designed to be as transparent and accessible as possible. "For each model it includes, it provides an explanation of the methodology it applies that is geared toward a general audience," he said.
The project was taken on independently by VMASC as part of its mission to develop computational models that help solutions in complex areas. However, several local decisionmakers, including Jim Redick, director of Emergency Management for the City of Norfolk, have begun to take interest in the platform.
The model uses an array of existing data sets, including county-level information published and updated daily by the New York Times, data from the American Communities Survey (ACS) and information about hospitalization, intensive care and mortality rate from COVID-19 published by the Centers for Disease Control.
The ACS dataset includes age-range demographics at the county level, which can be correlated with the CDC hospitalizations dataset to estimate an outcome for each case.
"In general, counties where a larger portion of the population are older are at greater risk for more severe outcomes," said Lynch, a senior project scientist who leads the Data Analytics Working Group at VMASC.
In addition, data from Twitter is included in the analysis in affected counties. Posts with tweets that mention "fever" or "cough" are collected daily, then run through a machine-learning algorithm to classify only those tweets of individuals self-reporting symptoms.
The research platform allows anyone to explore, at a very local level, how the number of reported cases is increasing or decreasing in different areas of the state, based on different models. The forecasting models show that some communities, such as Norfolk, have had more success leveling off the reporting of new cases than other localities.
The researchers also hope that the platform can aid other COVID-19 modeling efforts, featuring a section that points to more than 130 publicly accessible external data and modeling resources.
Gore added that this platform is not intended to be the definitive, county-level COVID-19 forecast.
"This is a research platform for exploring models," he said. "We believe the information provided by the platforms can be useful to the public, but our goal is to explore and explain how different models make predictions and forecasts."
VMASC is an applied research center of ODU focusing on innovation, workforce development, and industry ecosystem engagement programs leading to digital transformation. VMASC's research staff of more than 70 research faculty, scientists, support professionals and students perform scientific research, develop computational models and create information-to-insight and digital engineering solutions in areas such as digital shipbuilding, spaceflight and autonomy, and digital health and health equity.
