2024 State of Gaming Report
On Monday, August 21, 2017, the United States experienced an awesome event: the first solar eclipse visible across the contiguous U.S. since 1918. Millions of Americans paused their daily routines to step outside and admire our moon as it passed in front of the sun. As the lunar shadow traced its path across the country, a digital shadow of online activity traced along with it. The audience of the eclipse, enchanted by events in the sky, was offline.
Although everyone within the 48 contiguous states had at least a view of a partial eclipse, the path of totality was a narrow strip just 70 miles wide. This path began in Oregon and passed from the Pacific Northwest to South Carolina, including through major cities such as Kansas City and St. Louis. Any fixed spot within this region experienced darkness of night for a brief 3 minutes.
Knowing with high precision when and where totality was occurring, it is possible to reconstruct the moon’s path across the U.S. over time. Mapping this path against Comscore’s geolocated data, we can visualize the characteristics of internet traffic during the eclipse. In the following series of heatmaps, we display several snapshots of internet traffic at zip-code granularity across the United States (red=high, blue=low) alongside the center of the path of the eclipse as it moves across the country.
These plots tell a clear story: internet traffic decreases along the path of the center of the eclipse, and this feature is most dramatic during totality.
Source: Comscore Data Science; August 21, 2017
It is well-known that desktop browsing behavior and mobile browsing behavior have very different characteristics. To explore this, we drill down into a single city over which the total eclipse travelled and categorize traffic as originating from either a desktop or mobile user. In the following plot, we display the traffic in Kansas City over the UTC day on which the eclipse occurred, split into the two streams.
Source: Comscore Data Science
There is a striking deviation here between the two traffic streams: during the eclipse, desktop traffic decreased significantly more than mobile traffic did. Intuitively, this makes sense - people going outside to view the spectacle left their desktop machines idle but may still have been browsing on their mobile devices.
Studying this event from the perspective of Comscore’s data sources has provided more than an interesting study on internet audience behavior. It has also allowed us a unique opportunity to run a series of sanity checks on our measurements and metadata.
In summary, eclipses have scientific value beyond the realm of astronomy. As demonstrated here, we used this extraordinary and rare spectacle as an opportunity to further our understanding of digital audience behavior and to test our data’s ability to expose such trends.