A recent Scientific Reports study utilized Twitter data on social mobility to identify the population that exhibited greater and less compliance to public health guidelines formulated during the coronavirus disease 2019 (COVID-19) pandemic. These guidelines were formulated and implemented to contain the COVID-19 pandemic caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2).
Study: Twitter social mobility data reveal demographic variations in social distancing practices during the COVID-19 pandemic. Image Credit: Fatmawati Achmad Zaenuri / Shutterstock
Background
Social distancing and isolation were the two most effective methods to prevent the transmission of SARS-CoV-2 infection. These interventions were crucial, particularly during the early phase of the pandemic, when other interventions were unavailable.
At the beginning of the pandemic, US public health officials requested that the general public avoid organizing and participating in large gatherings and follow social distancing. Several studies have shown mixed compliance with these recommendations, which impacted its effectiveness. Some factors that disturb compliance with social distancing recommendations are the housing environment, financial burden, and distrust of public officials.
Understanding the factors that govern adherence to social distancing rules in different populations is essential. This can be achieved by assessing online mobility data to help understand travel patterns. Mobility data can be easily obtained from GPS-enabled mobile phones that can be used to assess mobility patterns during the COVID-19 pandemic. One drawback with this data is that factors that induced reduction in mobility cannot be assessed. Twitter could be an alternative source of mobility data.
About the Study
During the study period, public tweets could be collected in real-time from Twitter’s Application Programming Interface (API) for free. However, this was changed to a paid service by Twitter in 2023. It must be noted that geotagging in tweets is allowed in a Twitter platform that provides location information in the tweet metadata. Previous studies have utilized automatic Twitter geolocation to understand various patterns. The current study utilized public Twitter posts that contain location data.
A key advantage of Twitter data is the availability of multiple data on a user. The collected data can be used to establish the correlation between a user’s mobility and their health behavior, race, age, political affiliations, and income. These demographic factors help identify which population is more inclined to adhere to social distancing.
This study used data from the Twitter Social Mobility Index Project that included public geotagged tweets from the US between January 1, 2019 and June 21, 2020. An index was computed using Twitter’s geotagged data and estimated the standard deviation across locations within each week. High standard deviation values reflected a high mobility.
Study Findings
Analysis of variance (ANOVA) tests were conducted to identify specific groups in the US that exhibited more significant reductions in mobility during the COVID-19 pandemic. Asian and Latin American ethnicities, males, older age groups, and Democrats were among those who exhibited significantly reduced mobility. A similar mobility pattern was also found in people from higher population density states. Consistent with the findings of this study, a previous study also reported reduced mobility in Asian and Black populations.
A significant gender difference in mobility patterns could be due to imbalanced age distribution in the study dataset. In this study cohort, a comparable number of males and females were present in the under-30 age group. However, this balance was disturbed among people above 30 years of age, where males were more than females.
The current study provided important insights into the interactions between different groups. For example, political affiliation exhibited a greater impact when interacting with age and gender. As expected, age is an important factor for mobility. Older people exhibited a significant reduction in mobility irrespective of their gender. However, compared to other ethnic groups, older Black people showed a lesser reduction in mobility.
The main focus of this study was to determine the actual behavior and not the willingness or attitude toward social distancing. Individuals often agree with the importance of social distancing but have to travel for professional or economic reasons. Corroborating this observation, a previous study revealed that wealthy areas had reduced mobility during the pandemic. This study observed that increased trust in government was associated with greater mobility reductions across age and ethnicity.
Limitations of Twitter Data
Future studies must consider the limitations linked to Twitter data while utilizing it for research purposes. This data enables the assessment of individual characteristics instead of a holistic view that could also shed light on other issues. The data lacks important information, such as socioeconomic data, which could be a vital confounding factor in determining why some groups have smaller reductions in mobility. Further, the data could not determine whether individuals increased mobility was associated with other precautions, such as wearing facemasks and using hand sanitizers frequently.
Taken together, the current study highlighted the value of geolocated Twitter data in understanding how people responded or adhered to public health recommendations during the COVID-19 pandemic.
COVID-19 increases chronic fatigue risk