Dynamic predictability and activity-location contexts in human mobility

Area of Science:

• Mobility studies
• Human behavior analysis
• Data science

Background:

• Human travel is generally regular and predictable, influenced by individual, social, and global factors like pandemics.
• Understanding these regularities is crucial for various applications, from urban planning to public health.

Purpose of the Study:

• To investigate how variations in individual mobility, termed predictability states, inform population-level travel regularities.
• To explore the potential for more nuanced mobility predictions by analyzing temporal, activity, and location data.

Main Methods:

• Analysis of individual-level mobility data, focusing on temporal, activity, and location variations.
• Identification and characterization of 'predictability states' within human travel behavior.

Main Results:

• Predictability states exhibit distinct contextual and activity signatures.
• Location contexts are particularly effective in estimating mobility patterns, even with low-resolution or missing data.

Conclusions:

• Individual mobility variations contain significant information about population-level travel regularities.
• The findings support a more nuanced approach to short-term and higher-order mobility prediction, leveraging contextual information.