Uber’s Ride with the Sun: Tracking the 2017 Solar Eclipse

How did U.S. riders use the Uber platform during the solar eclipse of Monday, August 21, 2017?

To find out, we collected trip requests of around 40 major U.S. cities (Figure 1) near the totality path and put our data visualization tools to the test. In this article, we showcase how the recent solar eclipse affected Uber ride requests by harnessing the visualization power of our deck.gl and Voyager data visualization platforms.

The heatmap in Figure 2, below, depicts travel requests on August 21 during an 8-hour window before and after the eclipse:

In this diagram, the x-axis denotes the 40 cities being mapped and the y-axis represents time of day in Pacific Standard Time. Ride request frequency relative to a normal (non-eclipse) Monday is indicated by color, with yellow to red representing lower than usual frequency and green to blue higher than usual frequency.

All 40 cities selected were on or near the totality path and had at least 90 percent obstruction (meaning, the amount of visible area of the sun’s disc covered by the Earth’s moon). In our research, we compared the number of requests to a normal day and found that before and after the eclipse, there was a more than average number of trips requests across almost all of these cities. In fact, cities like Nashville, TN,  Atlanta, GA, and St. Louis, MO exhibited a consistently high number of requests throughout the day except while the path of totality passed through them.

In Figure 2, we notice that in smaller towns like Salem, OR, Columbia, SC, and Carbondale, IL, showing a significant increase in trip requests compared to an average day. Since they are located along the totality line, these cities were popular locations to watch the eclipse.

Using our heatmap, we determined a trip request pattern depicted by the red strip that follows the time and location of the eclipse as it made its way across the path of totality from coast to coast, signifying a significant decrease in trip requests during the eclipse. When matched with the time of total eclipse on the y-axis of the heatmap in Figure 3, below, we noticed that the dots matched almost perfectly with the red strip.

As depicted in Figure 4, below, we hypothesize that many U.S. Uber riders along the path of totality ended their trips and stopped traveling to watch the total eclipse!

Figures 5 and 6, below, are area graphs depicting trip requests in Nashville, TN and Columbia, MO. We see user activity drop right around the time that the total eclipse was passing over these cities and pick up again immediately after the eclipse passed.

By using visualizations to map rider activity during extreme events such as the solar eclipse, Uber Engineering can facilitate a more seamless experience for our users regardless of external conditions.

Interested in using data visualization to power the future of transportation? Consider applying for a role on our team!

Shan He is a software engineer on Uber’s Data Visualization team.