There is evidence that people are slowly starting to return to the office but our Research analysts wanted to gauge the extent to which this is happening.
Using geolocation data, they tracked foot traffic in 28,000 US offices1 to develop a model that adapts physics gravity-mapping models to measure the proximity of places to people. It works by measuring the distance between each point in a metropolitan statistical area (MSA) to the time-weighted locations of all the people in that area. Our analysts call it “geo-gravity”: the places that have the strongest pull are those that are most convenient to the greatest amount of people.
As of mid-May 2021, the data suggest that the number of people entering offices was down about 57% from pre-COVID levels - a substantial rebound from April 2020, when traffic was down almost 85%.
The maps below show the pre-COVID focal points in various city centres in 2019, compare them with 2021, and illustrate the gravitational shift between the two periods.
The data suggest that the pre-COVID focal points in cities (the hotter areas in March 2019) mostly remain the same in 2021, even as traffic migrated outward.
Click on each city below to see the gravitational shift:
Leveraging alternative data, machine learning, and advanced empirical methods such as causal inference, Barclays’ Data Science and Investment Sciences teams work to provide institutional clients with next-level insights and actionable ideas.
The Data Science team focuses on developing data science methods and building the technology infrastructure to work with data at a large scale, while the Investment Sciences team integrates data science into the investment Research process across all asset classes. The team has specialty in analyzing disparate data sets from geolocation, to credit card transactions, to text and natural language, to jobs, to news and social media.
Clients benefit from data science-informed Research written in collaboration with a wide range of Research analysts, especially equity and credit fundamental analysts, as well as proprietary algorithms and products that are derived from alternative data.