A New App Gauges Level of Covid Exposure Risk

Individual demand for information is rising. JL 

Kathy Pretz reports in IEEE Spectrum:

A cloud-based smatphone app uses publicly available databases to tell when banks, grocery stores, parks, and pharmacies are less crowded. A number of computer-based  epidemiological models could predict the spread of COVID-19. Although those models have been good for health departments, “they do not provide individuals—on a moment’s notice—[information on] how to avoid getting infected.The app uses Google Maps popular times option, which shows how busy a place tends to be at specific times. The data is combined with information about COVID-19 infection rates, vaccinations reported, and surveys about the willingness of people to wear a face covering, all sorted by ZIP code.

As COVID-19 mandates are lifted, those who are immunocompromised or elderly will be looking to protect themselves from coronavirus exposure by avoiding public places that are packed with people.

Albert Cheng, a professor of computer science and electrical and computer engineering at the University of Houston, is developing a smartphone app to help them do that. The cloud-based app uses publicly available databases to tell when banks, grocery stores, parks, and pharmacies are less crowded. He is currently piloting the app in Houston and Seattle.

Cheng unveiled his app at the 2021 IEEE/ACM High Performance Computing Conference for Urgent Decision Making, sponsored by the IEEE Computer Society. His “Real-Time COVID-19 Infection Risk Assessment and Mitigation Based on Public-Domain Data” research paper is available in the IEEE Xplore Digital Library.

THE DEVELOPMENT PROCESS

Before Cheng started his project, there were already a number of computer and epidemiological models that could predict the spread of COVID-19, he says.

Although those models have been good for health departments and government officials, he says, “they do not provide individuals—on a moment’s notice—[information on] how to avoid getting infected if they need to do grocery shopping, banking, or other activities.

“I’d like to empower them,” he says, “with up-to-date information, all in one place.”

Cheng’s app uses Google Maps’ popular-times option, which shows how busy a place tends to be at specific times during the day. The data is combined with information from public health agencies about COVID-19 infection rates, the number of vaccinations reported, and surveys about the willingness of people to wear a face covering, all sorted by ZIP code.

“All these parameters are input into an algorithm that weighs the information to determine the risk for infection at different stores near the user,” he says. “The more people who are vaccinated, the lower the risk.”

The app recommends which store in the ZIP code has the lowest infection rate as well as the best time of day to visit it. He says users might find that a popular store presents a lower risk for catching COVID-19 than one that is less crowded, because the shoppers there are more likely to be vaccinated and to wear a mask.

Users also can search by the name of a specific place and get current traffic conditions. Cheng is working on giving users the ability to select locations within a specific radius.

For this store, the app displays its address, the times when the store is less populated, and the probable number of COVID-19 cases in that ZIP code.ALBERT CHENG

The app was supported with a grant from Cheng’s university. He says the school plans to license the technology to his new AMKC Informatics startup. He’ll be running the company while continuing to teach and research real-time systems and embedded systems, he says, adding that his expertise in scheduling and routing will come in handy.

DATA HURDLES

Developing the app was no easy task, he says. States—and even counties within the same state—use different formats to track data about COVID-19 infections and vaccinations. Some state health officials archive their historical data; others delete it. Some agencies release updates daily; others less frequently. In the Houston area, for example, Harris County’s records are updated daily, but only on weekdays. Florida is considering releasing its report only once per week, Cheng says.

“The less data we have, the [more the] app’s accuracy suffers,” he says. “We must use more extrapolation techniques to make inferences with the limited amount of data that’s publicly available.”

The app will be offered in more cities, he says, including some outside the United States, once he figures out how to address the different data formats.

“It's a very time-consuming process to do that,” he says. “Right now, the app cannot serve all cities.”

MAKING AN IMPACT

Even after the pandemic is declared over, the app will be useful, he says, because variants will continue to emerge. The app also could assess the risk of catching other viruses including influenza.

“It has always been my goal to not just publish papers and get funding but to also translate research into something that positively impacts society,” Cheng says. “I think this is a juncture in my career where there’s a very good opportunity for me to make an impact and provide a service that reduces the chance of infection, reduces the number of positive cases, and basically makes people healthier.”