DestinationsHow an email to the prime minister and algorithms helped Greece bring tourists back.

Reopening to tourism? It's all Greek to me

By
|
A data science professor's algorithms were estimated to have caught twice as many positive cases of the virus than would otherwise have been with Greece's limited resources.
A data science professor's algorithms were estimated to have caught twice as many positive cases of the virus than would otherwise have been with Greece's limited resources. Photo Credit: Getty Images/Maglara

Greece often triples its population through annual tourists so opening up safely to visitors in the summer was crucial.

While the country relies on tourism for 20% of its GDP, testing resources were scarce back in May when it began to address the challenges of welcoming tourists during the pandemic.

Watching from afar was Kimon Drakopoulos, assistant professor of data science at USC Marshall, who had looked at mathematical models for understanding pandemics in the past.

Curious after reading about Greece’s plan to reopen, he decided to email the country’s prime minister and several hours later received an unexpected response.

From that email in May, with a team of four working on a solution 24/7, Project EVA prototype went live on 1 July.

While the rest is history, albeit recent, Greece reopened with the help of algorithms to track worldwide infections and figure out who to test from the millions of visitors coming to the country.

The algorithms looked at where people were coming from, places they had recently visited, when they were arriving and general demographic information to ascertain whether they could be in the group of people who are infected but asymptomatic.

While it might sound simple, it was a huge undertaking with more than 40 points of entry to be taken into consideration, 50,000 visitors a day, 8000 tests in the laboratory, teams of doctors and nurses and transport logistics all in the mix.

The results were very positive with the algorithms estimated to have caught twice as many positive cases of the virus than would otherwise have been with the limited resources.

However, it also had much wider influence in terms of good policy decisions in terms of screening protocols and prohibiting visitors from some countries.

While the summer season was slower as expected for tourism to the country, Greece still managed to bring back 25% compared with the 2019 season, against an expectation of between 10% and 15%.

It also managed to prolong the season to October by controlling the borders and staying on top of the virus.

For more on Project Eva, how tourists reacted and the challenges of bias in algorithms, watch the full discussion between Geoffrey Garret, dean, USC Marshall, Drakopoulos and Vishal Gupta, assistant professor of data science and operations, USC Marshall.

Source: Phocuswire

JDS Travel News JDS Viewpoints JDS Africa/MI