Exploring daily rhythms of interpersonal contacts: Time-of-day dynamics of human interactions with latent class cluster analysis

This study used activity-travel diary data to analyze the time-of-day dynamics of interpersonal contacts and examined their complex relationship with other activity-travel time allocation and personal accessibility dynamics. In total, 2,942 activity-travel diaries from 1,471 participants were used to identify five unique patterns of daily human interaction, with latent class analysis. Latent classes for the time-of-day dynamics of time allocation to activities and the time-of-day dynamics of experienced business employment density are also estimated independently and correlated with the five human interaction patterns. The analysis used a form of multinomial regression model (also called a three-step model in latent class analysis) to examine these relationships and to test the association of human interaction patterns and external explanatory variables simultaneously. Strong correlation was found between interpersonal contact patterns and activity participation patterns, along with day of the week, gender, and age. Other sociodemographic indicators and business employment density only partially explain these dynamics.