How the Subway Warned Me of the Pandemic, and What It’s Telling Me Now

The earliest omen that I noticed of the pandemic’s potential impact was the increasing frequency of Asian tourists (1) wearing masks on NYC’s subways, starting around mid-to-late January. [Incidentally, stock prices in Asian businesses that manufacture rubber gloves rose around this time as well (2)]. Though Asian cultures have a history of mask wearing (3), most Asian tourists whom I’ve seen in the past followed the customs of New Yorkers and did not wear masks during their visits.  

Due to my more frequent subway sightings of masked tourists/love of reading the news, I may have started worrying a little earlier than most U.S. residents about the possibility of a pandemic. At the very least, on January 23^rd, I urged my family to get masks; however, they were not convinced of the severity of the situation. At the time, there were only 10 confirmed cases of COVID-19 outside of China (4) and the World Health Organization “could not reach a consensus” as to “whether the outbreak constituted a public health emergency of international concern” (5). Coronavirus was first detected in NYC over a month later, on March 1st, after a healthcare worker returned from a trip on February 25^th (6).  

I knew that we were heading towards serious trouble in late February-early March when I saw two different people—who were clearly not the average tourists based on their work attire—wearing a mask and disposable rubber gloves during early morning rush hour. The second sighting stood out to me because I mentioned this to an acquaintance on March 2^nd and they dismissed that person as “crazy”. During the week of March 13^th, when President Trump declared a national emergency, disposable rubber gloves littered my neighborhood. Non-essential businesses in NY were shutdown on March 22nd (7).

Subway Case Study: Background

Now that we’re in August, it’s rarer to see people wearing rubber gloves, but New Yorkers will look at you askance if you do not wear a mask. This of course, is just my anecdotal first-hand experience. But as NYC begins to reopen and because my observations on the subway played such a large role in alerting me to the potential impact of COVID-19, I wondered if I could use actual data to link the subway to the closing and re-opening of businesses in NYC.  

This is not a new concept. Financial firms have been using “alternative data” for years now to help give them an edge. Alternative data is the use of non-traditional data, such as social media or satellite imagery, to help inform financial decisions. For instance, instead of waiting on the Bureau of Labor Statistics to release traditional unemployment numbers, a firm might estimate unemployment by scraping the web to look at the number of job postings/applicants.  

Method:

I built a web scraper in python to download data from the MTA. I sorted through ~37 million rows and focused on the most popular eight out of 378 subway stations across Manhattan, the Bronx, Brooklyn, and Queens. I graphed the data in R. The subway stations show similar trends in ridership. For this article, I decided to share my graph based on the most-utilized subway station in 2019: Time Square 42nd Street (8).

In the graph (9) below, the blue line represents subway ridership in 2019, and the orange line reflects ridership in 2020.

Analysis:

The regular “jumps” in the lines represent increases in ridership during the weekdays and the declines represent decreases in ridership during the weekends. The consistent “jumps” in subway usage during weekdays suggests that the subway is a popular means of commuting to work. Therefore, the subway may be an especially valuable indicator for evaluating how the nature of work has changed.

Subway ridership at the Time Square 42^nd Street Station in January 2020 was lower than in January 2019. Ridership in 2020 starts decelerating even faster in February, weeks before a State of National Emergency was declared on March 13^th, 2020. It seems that other people were similarly nervous about the possibility of a pandemic and took steps to minimize exposure even before the government’s declaration. This suggests varying confidence in the government’s guidelines on COVID-19, which will likely hamper economic recovery. If people are skeptical of the government’s guidelines on when it’s safe for businesses to close and reopen, businesses may not re-open in a timely fashion. Notably, even the declaration of Phase 4 of NY’s reopening did not result in as large of a change in ridership compared to the declaration of a National Emergency and NY’s Stay-at-Home order.

Just as President Trump’s March 13^th declaration of a national emergency, local government guidance under NY’s Stay-at-Home Order led to a visible, sharp decline in subway ridership. However, subway ridership began to slowly increase even before the NY State Government initiated Phase 1 of reopening. Although NY is now in Phase 4 of reopening, subway ridership is clearly nowhere close to its previous levels. Ridership should continue to grow as businesses reopen.

 At the start of the pandemic, there was much discussion about whether the U.S.’s economy would have a “V-shaped” recovery (10). This would occur if the economy’s sharp downturn was matched by an equally sharp upturn.

My graph of subway ridership suggests that the economic recovery will be shaped more like a “swoosh”, with a large drop in activity followed by a slower recovery. This is consistent with guidance from companies in sectors such as consumer goods, airlines, and restaurants (11). Even with the lifting of lockdowns, large gatherings are still prohibited and retailers are limiting the number of customers allowed in stores. Hence, economic activity is resuming, but is not “springing back” as in a V-shape.  

Looking Forward:

The trend in subway ridership lends itself to interesting revelations; however, the subway may not be the best way to track NYC’s recovery from COVID-19 in the future. If remote work becomes the “new normal”, subway ridership may never return to previous levels. Other sources of alternative data may be more helpful. 

I recommend browsing TracktheRecovery.org, a site created in partnership with organizations including Harvard University, Brown University, and the Bill and Melinda Gates foundation. This website has other indicators such as consumer spending, small business revenue, and low-income unemployment. Like subway ridership, these indicators suggest a trend in economic activity consistent with a “swoosh” shaped economic recovery. It will be interesting to see if this swoosh shape holds in the future.  

Notes/Sources:

1. I assumed that any individual whom I saw rolling a suitcase or consulting map were “tourists”, although I acknowledge this is not always an accurate assumption.
2. Financial Times, “Coronavirus outbreak boosts rubber gloves and surgical shares”, 2020-01-22.
3. BBC, “Coronavirus: Why some countries wear face masks and others don’t”, 2020-05-12.
4. World Health Organization, “Novel Coronavirus (2019-nCoV) SITUATION REPORT – 3″, 2020-01-23.  
5. World Health Organization, “Archived: WHO Timeline – COVID-19″, 2020-04-27.
6. Wall Street Journal, “First Case of Coronavirus Confirmed in New York State”, 2020-03-01.  
7. For more information on NY’s shutdown and reopening, please see: https://coronavirus.health.ny.gov/new-york-state-pause and https://forward.ny.gov.
8. MTA, “Subway and bus ridership for 2019”, accessed on 2020-08-13.
9. The January 2019-August 7th, 2020 data is current as of August 8th, 2020. I created this image using MTA data.
10. CNBC, “Why the chances for a ‘V’-shaped economic recovery are getting less likely by the day”, 2020-04-06.
11. Wall Street Journal, “Why the Economic Recovery Will Be More of a ‘Swoosh’ Than V-Shaped”, 2020-05-11.
12. TracktheRecovery.org. Image accessed on 2020-08-13.