[Paper Review] Mobility Changes in Response to COVID-19

Michael S. Warren, Samuel W. Skillman|arXiv (Cornell University)|Mar 31, 2020

Human Mobility and Location-Based Analysis11 references132 citations

TL;DR

The paper analyzes anonymized mobile device location data to quantify global and US mobility reductions due to the COVID-19 pandemic and related policies, and releases admin1/admin2 mobility data under CC BY 4.0.

ABSTRACT

In response to the COVID-19 pandemic, both voluntary changes in behavior and administrative restrictions on human interactions have occurred. These actions are intended to reduce the transmission rate of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). We use anonymized and/or de-identified mobile device locations to measure mobility, a statistic representing the distance a typical member of a given population moves in a day. Results indicate that a large reduction in mobility has taken place, both in the US and globally. In the United States, large mobility reductions have been detected associated with the onset of the COVID-19 threat and specific government directives. Mobility data at the US admin1 (state) and admin2 (county) level have been made freely available under a Creative Commons Attribution (CC BY 4.0) license via the GitHub repository https://github.com/descarteslabs/DL-COVID-19/

Motivation & Objective

• Assess how voluntary behavior changes and government restrictions affected daily human mobility during COVID-19.
• Develop metrics to quantify mobility from mobile location data.
• Make mobility statistics available publicly for policy analysis and epidemiological modeling.

Proposed method

• Process large-scale mobile device location data with cloud computing and the festivus file system to compute mobility statistics.
• Filter reports by position accuracy and collate per-node daily reports.
• Compute three mobility measures: M_max, M_bb, M_ch, and focus on the median M_max (m50).
• Define a normalized mobility index m50_index using a regional normal m50 based on a pre-COVID baseline period.
• Reverse geocode canonical locations to country and admin regions (Admin1, Admin2) for aggregation.
• Output results in NDJSON and CSV formats for streaming analysis.

Experimental results

Research questions

• RQ1 How did mobility change in response to the onset of COVID-19 and subsequent policy measures at global and US subnational levels?
• RQ2 How do mobility metrics (M_max, M_bb, M_ch) reflect changes in daily movement patterns during the pandemic?
• RQ3 What is the effect of shelter-in-place and other policies on regional mobility as quantified by m50_index?
• RQ4 What are the limitations and potential biases in using anonymized mobile location data to estimate population mobility?

Key findings

• Mobility reductions occurred globally and in the US following the COVID-19 threat and government directives.
• In the US, mobility dropped markedly with New York falling from 5.2 km to 31 meters on certain dates, indicating most people stayed near their initial location.
• Normalized mobility index shows Florida and Texas down to ~30% of normal, and California, Illinois, New York, and Washington to <20% of normal during early March 2020.
• Significant county-level mobility changes correspond to local events such as Bay Area shelter-in-place orders and university campus closures.
• Dramatic mobility changes aligned with major policy actions and holidays; trends varied by rural/urban differences and by local circumstances.