Exploring Invariants & Patterns in Human Commute time

TL;DR

This study analyzes recent commute patterns in two Chinese cities, revealing slight increases in commute times and changes in travel behavior, challenging the longstanding Marchetti's constant.

Contribution

It provides new insights into how urban development affects human commute time patterns using fine-grained communication data.

Findings

Commute time has slightly increased over recent years.

Medium and long-distance commuters' endurance limits have grown.

Population distribution across commute distances remains similar across cities and over time.

Abstract

In everyday life, the process of commuting to work from home happens every now and then. And the research of commute characteristics is useful for urban function planning. For humans, the commute of an individual seems revealing no regular universal patterns, but it is true that people try to find a satisfactory state of life regarding commute issues. Commute time and distance are most important indicators to measure the degree of this satisfaction. Marchetti states a certain regularity in human commute time distribution - specifically, it states that no matter when, where and how far away people live, they always tend to spend approximately the same average time for their daily commute. However, will the rapid development of cities nowadays as well as serious challenges brought by economic development affect this constant? If there are novel characteristics? We revisit these problems using fine grained communication data in two Chinese major cities during recent two years. The results indicate that the commute time has been slightly increased from Marchetti's constant with the development of society. People's overall travel budgets have been increased, more concretely speaking, for medium and long distance commuters, their endurance limit for commute time is enhanced during the passing years, and fluctuates around a constant; for short distance commuters, their commute time increases with the distance. Moreover, the population distribution in every commute distance shows strong cross-city similarity and does not change much over two years.