- Part of our Aggie Innovation Platform series

The COVID-19 pandemic has drastically affected both lives and livelihoods — impacting communities and disrupting economic activities throughout the world.

Dr. Ali Mostafavi, associate professor in the Zachry Department of Civil and Environmental Engineering, and his Urban Resilience.AI Lab, developed a deep-learning model that can predict the spread of COVID-19 cases for U.S. counties with 64% accuracy — twice the accuracy of an untrained model. 

Mostafavi is using data-driven research methods combined with cloud technology in the Division of Information Technology’s Aggie Innovation Platform (AIP) to combat the impact of the pandemic.

Deep Learning Model Predicts COVID Surges

A deep-learning model is a type of artificial intelligence where computing systems learn from large amounts of data. By training the deep-learning model with data from a certain time period, the model identifies features it can use to predict the trajectories of another time period.

Mostafavi’s team collected a vast amount of data, including census data, social demographics, social  distancing statistics, case-count growth and population movement within the community.

“This model doesn’t identify specific mitigation or response strategies, but it helps highlight which strategies could be effective based on various county-level data,” Mostafavi says.

When communities can forecast the spread of COVID-19, they can identify the best way to mitigate the surge. For example, if mobility is the most critical feature for a county, officials can establish stay-at-home policies.

Cloud + IT Partnership Empowers Innovation

This crucial research initiative is one of the first Texas A&M research projects to utilize the Division of IT’s Aggie Innovation Platform (AIP) cloud technology.

Mostafavi began using the cloud for research in 2017 after Hurricane Harvey. For his research on prediction and response to urban disaster situations, Mostafavi collected data on traffic changes, satellite imagery, social media, flood impacts — and all this data needed to be stored securely with sufficient CPUs to process.

“My lab had a server, but not one with enough storage or computing capacity,” states Mostafavi. “I applied for a grant from Amazon Web Services (AWS) and moved my research to the cloud.”

When the pandemic hit in 2020, Mostafavi’s research focus changed, and his data use once again exploded.

“I went from 20TB to 150TB in just three months. I knew I needed a partner to help me make the best use of my cloud infrastructure. By partnering with AIP and Engineering IT I was able to make full use of all the resources available to me. They were tremendously helpful — I didn’t even know what I didn’t know. “ 

What Does This Mean for the COVID-19 Pandemic?

Current results show that big data and deep learning have the potential to track and contain disease spread, as well as inform leaders about the best ways to mitigate the impacts of that spread.

Going forward, the Urban Resilience research team will use new datasets to develop new types of models. The group is currently working on an AI-based model for city-scale surveillance to predict cases at the zip code level. The goal will be to predict the factors that influence each zip code so officials can explore location-specific policies.

Overall, Mostafavi is optimistic. “Significant opportunities exist using big data and AI to contain the existing pandemic and also better prepare for and mitigate future pandemics.”