An interactive tool will let you calculate the chance that someone at your Thanksgiving dinner may have COVID-19

• Researchers created an interactive tool that calculates your chances of encountering a person with COVID-19 based on your location and how many people are at a gathering.
• The tool shows a county-by-county breakdown of your coronavirus risk in real-time.

With Thanksgiving less than two weeks away, Americans are contemplating how much risk they're willing to take on in order to hang out with loved ones.

To help with that assessment, researchers at Georgia Tech created an interactive tool that calculates the risk for you.

A study published Monday describes the tool, which can tell you the likelihood that you'll encounter at least one person infected with COVID-19 at a gathering, depending on your location and the size of the group.

According to Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, indoor gatherings are driving the unprecedented coronavirus surge in the US. So ideally, people should avoid close contact with those outside their households to lower their risk of getting the coronavirus.

But if you are planning to see your friends and family this holiday season anyway, the risk-assessment tool can offer a sense of just how big a risk you're taking. The numbers get updated in real time, so your risk may change from day to day.

"It gives people a way of envisioning themselves in a situation and be able to decide whether to go to the event or not," Clio Andris, a professor of city planning at Georgia Tech and co-author of the new study, told Business Insider.

Say you are planning a dinner party with nine other guests on Friday. If you're in New York City, the tool shows that you'd have at least a 12% chance of sharing air with an infected person. But if the party is in Chicago with a group the same size, the risk jumps to more than 50%.

Risk depends on the gathering size and your county

To use the tool, first select how big your gathering is — between 10 and 5,000 people. Then you'll be prompted to hover your cursor over the US county where your gathering will take place.

The probability that an infected person will be present is shown on a scale from 0% to 100%. The data underpinning the math is updated everyday, Andris said, based on daily case numbers collected by The New York Times.

The tool shows how different the risk can be from county to county: The higher the number of cases per capita, the higher the risk. Risk also grows as groups get larger.

As of Friday, for instance, there's an 89% chance that an infected person will be at a 10-person gathering in Norton County, Kansas. Compare that to 11% for the same size party in Monroe County, New York.

The tool also attempts to factor in the asymptomatic carriers that COVID-19 testing might be missing. According to Andris, the researchers introduced an element into their calculations called "ascertainment bias," which is an estimate of the number of actual coronavirus cases per documented case. They based this estimate on July research that compared the total number of people who tested positive to the total number of people with antibodies (which indicates they'd previously been infected) at 10 sites across the US.

The tool offers two "ascertainment bias" options: Users can have the tool assume that the actual number of coronavirus cases in the US is either five or 10 times higher than official tallies. That latter option is likely more accurate, according to an August report by the Centers for Disease Control and Prevention.

A larger ascertainment bias also results in increased risk.

A 'decision support system'

What the tool cannot do, the study authors said, is tell a user the likelihood that they might get infected at a gathering. Rather, it's designed as a "decision support system that allows individuals to measure the risk of their own actions and plan accordingly," they wrote. `

Andris helped the tool's creator, Joshua Weitz, get the dashboard up and running in July. Weitz's inspiration came from his desire to determine the probability that someone infected with COVID-19 would have been at an Atlanta United FC Soccer game this winter, before the team played their last game on March 11.

By September, more than 2 million people had visited the site.

Andris even used it herself to decide whether to visit three friends near her home in Fulton County, Georgia, on election night. The tool said the risk at that time was 3%.

"It was my first time indoors with friends since March," Andris said, adding, "the risk was minimal, but it's not something I would do every night or every week."

In fact, she said, Andris is likely planning a "lay-low type of Thanksgiving" this year, even though she hasn't seen her family in Washington DC in eight months.

She hopes other holiday travelers will use the tool to inform their decisions, too.

"We know it's been a hard year for people, we really want to get together and see each other," Andris said. "But we hope people use this tool to reassess their plans."