Study to set groundwork to build machine learning models that rapidly analyze how a virus spreads
In April when New York City was under a strict lockdown, a team of 16 student researchers from New York University’s Tandon School of Engineering commenced a National Science Foundation Rapid Response Research (RAPID) grant-funded project to observe potential COVID-19 hotspots outside of hospitals and mass transit hubs. Utilizing virus mapping techniques that date back to 1854 and more modern, geospatial data, the researchers recorded what people touched, identifying surfaces most likely to carry the coronavirus. The project was driven and managed by Debra Laefer, a professor of civil and urban engineering at NYU Tandon.
Collectively, the team spent over 1,500 hours recording observations across all hours of the day and days of the week. They collected 5,065 records documenting the behavior of 6,075 individuals around 19 hospitals and urgent care clinics across four of New York City’s five boroughs. Observing subjects from a distance for periods of up to 20 minutes, the researchers documented the gender of the subject, objects touched, route taken, and their next destination. Each record also included locational information about the facility, the weather, and 61 other attributes related to the demographics of the facility’s zip code.
The researchers discovered that:
- 75% of individuals touched something after leaving the building
- 11% touched their phone
- 55% left the area by a form of mechanized transportation
- 13% returned to the medical facility
- 81% were wearing PPE
The team has made the complete data set available as a resource for scientists building machine learning models to map and analyze the spread of the coronavirus. Download at NYU’s Spatial Data Repository.
“As soon as COVID-19 started impacting New York, I sought to find a way to use my skills to help further research and many of our students felt the same way. After receiving the grant from the National Science Foundation, we were quick to organize and start collecting data,” said Laefer. “While it was great to see such a high percentage of people wearing PPE, many were quick to touch their phone or other objects, potentially spreading the virus. We hope this data will help create more accurate virus tracking models.”
Learn more about the New York University Tandon School of Engineering here