Follow Datanami:
May 18, 2020

New Dashboard Shows How COVID-19 is Changing Foot Traffic for Top Brands

It’s no secret that for most of the world, travel habits have changed significantly over the last few months. As some areas continue to enforce strict lockdowns and others cautiously start to reopen their doors, more scrutiny than ever is being paid to how people are traveling within their communities. Now, a new dashboard from Gravy Analytics is using real location data to make that picture even clearer for researchers and decision-makers, showing how American consumers’ travel habits are changing as the global crisis evolves.

Gravy Analytics was founded in 2011 just outside Washington, D.C. Since then, the firm has been crunching away at massive amounts of data – tens of billions of location signals – to understand trends in human mobility and consumer behavior. “When the COVID crisis hit, we understood that obviously human mobility had a large part to play,” said Jeff White, founder and CEO of Gravy Analytics, in an interview with Datanami. “Not only in understanding the virus’ progression, but ultimately in how we cope with it, how we leverage it to save lives and ultimately get back to some level of normalcy as quick as we can.”

So the team at Gravy built a dashboard – not a dashboard tracking COVID cases or hospital beds, but a dashboard tracking movement. Gravy’s dashboard charts the trends of in-person visits to 200 different retail and restaurant chains across the country, setting it apart from tools like Apple’s Mobility Trends tool which seek to show movement trends during the pandemic by displaying overall changes in routing requests.

The data in the dashboard is broken down by either state or major city (50 are available, for now), and can be viewed either in terms of specific brands (e.g. AMC Theatres or Petco) or categories (e.g. movie theaters or pet stores). The dashboard reaches back to early February and is regularly updated.

Weekly change in daily visitors to tracked locations by state as of May 3, 2020.

“You start with accuracy and fidelity,” White said. Gravy, he explained, painstakingly ensures that each logged visit is precise and reliable. To do this, they examine both the anonymized GPS data and 10 million hand-drawn polygons that Gravy uses to encapsulate each business location, serving as thresholds for consumer visits. “We have a bunch of metadata algorithms that validate those attendances – so we’re not doing any modeling, we’re not doing any inference-based analytics, it’s 100% deterministic, leveraging both of those data verticals.”

“We see, at the top of the funnel, 20 to 30 billion location records every day,” he continued. “By the time we cleanse, validate, throw out the garbage, and do all these other things, we’re down to about half of that that is validated and verifiable in our platform.”

The data, when viewed through the lens of the coronavirus pandemic (and processed through Tableau, which Gravy used to create its visual dashboard), can reveal some fascinating truths. Some are more predictable (did you know that movie theater attendance has gone down?); others, a bit less (did you know that gym attendance has gone down even more?). 

Daily changes (national) in visitations to Sonic Drive-In over the course of the pandemic.

“You see the hoarding impact of the lockdown,” White said. “You’ll see a normalized dataset, and all of a sudden right around mid-March, it spikes – everybody’s buying all the toilet paper, going to all the grocery stores, all those things – and then it leveled out, right? So our visitations dropped way down.” White also said he was curious why Sonic, in particular, seemed to have relatively normal visitations compared to other quick service restaurants: “And then I realized, Sonic, well, they’re a drive-in model, they always have been!”

Adapting Gravy’s analytics process for a pandemic-stricken world required some quick thinking on the team’s part. As stores and restaurants began to do more and more curbside pickup, Gravy’s tightly-bounded polygons started capturing fewer and fewer of customers’ visits to those locations. “Very early on, we started to notice that behavior,” White said. “So we quickly shifted our polygons to include the curbside pickup locations across those major brands, where appropriate – so we got a much better accuracy ready on those.” As it turns out, curbside pickup became a boon for Gravy in some cases, with longer visits at many locations ensuring greater accuracy of the GPS readings.

Gravy is now creating a wide range of COVID-19-related tools. “We’re rolling them out all the time,” White said. “We understand that the automotive plants are getting ready to come back online, so we’re rolling out category-level dashboards now. We’re going to do this at a much more frequent pace – and then we’ve also made several datasets free on [the AWS Data Exchange] for those that want to … put them in their own analyses and their own tools.”

White is also looking ahead to different applications of Gravy’s location data as the pandemic continues and mitigation measures ramp up. “I think as more and more of the effort gets to contact tracing and other things,” he said, “this data has a role to play there.”

To access all of Gravy Analytics’ COVID-19 tools, click here. The state-level dashboard is available here and the city-level dashboard is available here.