Using Population Movement Analytics to Battle COVID-19 Spread

This is an important message to the engineering.com community. Messages that are informative, helpful and relevant to COVID-19 will be posted here as a special service during our time of crisis. Messages are not written by engineering.com. If you have a COVID-19 story or message, please submit here.

By Aaron Williams & Jonathan Wolf 

COVID-19 is more than an outbreak; it’s a journey. Authorities and business leaders are faced with a multitude of decisions as the pandemic works its way through humanity—each with its own powerful consequences.

What these decision-makers, from public health and elected officials to business executives and public safety leaders, need most is the ability to comprehend and monitor what’s happening in their communities, in real time. Understanding population movement, patterns of congregation, workforce dispersal and more, in numbers by the thousands, millions or even hundreds of millions, is essential to making good policy decisions both now and in the future.

Specialized places and patterns data, presented within a rich analytics environment that covers both space and time, and supports interrogation at the speed of curiosity, will make all the difference in the next phases of the COVID-19 fight. It’s a capability that OmniSci and SafeGraph, as partners, have created.

Key Population Insight

OmniSci and SafeGraph joined their technologies to provide insight into movement patterns at millions of public and commercial locations.

If a place has a name, anywhere in the United States, SafeGraph has data about people who go there, constantly updated. The company is considered the leader for population movement and volumetric data at POIs (points-of-interest), including commercial and retail businesses, and public spaces. SafeGraph quantifies visitor foot traffic, demographics, where people are coming from and, often, where they’re going next. Moreover, it collects that information over time to uncover trends and tendencies.

To address COVID-19, SafeGraph created a new data set called Social Distancing Metrics. Updated daily, it looks at neighborhood-by-neighborhood and county-by-county data to understand relative volumes of movement. Information on five million POIs, 5,500 retail chains, and three million small businesses across 60 categories are included in the data set. Also incorporated is accurate information on the percentage of the population sheltering in place each day.

OmniSci, a pioneer in accelerated data analytics through the power of parallel processing, has the ability to analyze and geographically map the billions of rows of data in SafeGraph’s data stores in milliseconds. The speed and scale of OmniSci’s capabilities allows users to interact with these massive amounts of data in order to gain spatiotemporal insight as quickly as questions can be posed.

Plotted geographically through OmniSci’s Immerse web-visualization interface, SafeGraph data unlocks answers to many of the most difficult questions related to the pandemic.. This capability is of value to policymakers, academics and business leaders across the spectrum of organizations impacted by COVID-19.

How It’s Applied

Many questions critical to the pandemic are about movement. Retailers and business leaders need to identify how to modify behavior among employees and customers; by analyzing how aggressively people are engaging in their communities, for example, and comparing that with the rise of COVID-19 in that area, they can get a sense of how much awareness they need to build, or what policies to enact. 

How much education is necessary? What do employers need to focus on? For restaurants and retail businesses, how effective are distancing policies? As a manufacturer, how do I ensure my workforce can operate safely at full strength? Knowing how groups of people in a particular area are gathering or traveling can make these difficult decisions much easier.

OmniSci, through its Immerse interface, allows users to plot and view granular population movement data, as well as unaggregated visitation data. It supports many interrogative tools, from standard visualizations such as line, bar, and pie charts to more complex visualizations including geo-point maps, geo heat maps, choropleths and scatter plots. 

It also creates the ability to see changes over time. By looking at the current situation versus pre-COVID, for instance, it’s possible to see the impact of recent decisions. Users can see where people are congregating on weekdays versus weekends, or how worker movement differs from those who work at home.

Predictive analytics is another valuable capability. One of the key prerequisites of predictive analytics is a significant amount of data, something that SafeGraph collects. Forensic data supports good decision-making—but forensic plus predictive data allows even better decision-making. Using machine learning algorithms, OmniSci supports logistic regression, decision trees, and time series analysis to model the possible outcomes of different policy alternatives.

(Demonstration courtesy of SafeGraph and OmniSci.)

Broad Uses 

The combination of SafeGraph data and OmniSci horsepower lets policymakers understand the correlation between a rise in cases and population movement in certain types of businesses. Recent issues at U.S. meat processing plants underscore the need for population movement analysis. Several plants across the country were forced to close because of coronavirus spread between workers, and these infections had significant impact in the surrounding regions.

Authorities can get employees back into their work environments safely when equipped with a better understanding of what is happening on a community-wide basis. The data also holds value to other communities within similar industries that want to be prepared for potential outbreaks. By being able to use both good and bad benchmarks in public policy and commercial enterprise, better, more informed decisions can be made.

At the retail level, analysts may correlate restaurants that are reopened for dine-in against a rise in coronavirus cases, versus those offering options other than dine-in. This can be immensely helpful for individual restaurant owners as well as public health authorities as they answer questions about how to service customers, how to arrange seating, and how quickly to expand their dining options.

Because SafeGraph covers virtually all POIs and places of business, it’s possible to compare one category against another; e.g., essential businesses versus non-essential, grocery stores versus public markets, or one store location versus another. Decisionmakers can also see the impact of group gatherings at party centers, country clubs, recreation centers, athletic fields, amusement parks or family fun centers.

Finally, comparing SafeGraph data with other public or private data sets within OmniSci can add further insight. COVID-19 case counts, customer point-of-sale data, weather statistics, and other third-party data enriches the decision-making environment. By plotting these forms of data against population movement, it’s possible to create detailed plans that take into account diverse and even unexpected influences.

Free to Academics, Non-Profits

OmniSci and SafeGraph are making their offerings available at no cost to academics doing research related to COVID-19 response. Specifications and details of these capabilities can be found on websites for both OmniSci and SafeGraph.

Businesspeople, academics and public policymakers know that this story isn’t going to end anytime soon. There will be important decisions to wrestle with for the next year or more, each with its own challenges and considerations. The journey to our post-COVID world may be a difficult one—but with the right data and the analytics in the right hands, the road ahead may become a lot straighter.