How many days are enough?: capturing routine human mobility.

Area of Science:

• Geographic Information Science
• Human Mobility Studies
• Spatial Behavior Analysis

Background:

• Mobile phone sensor technology, particularly Global Positioning System (GPS), offers high-fidelity human mobility data.
• Existing data collection methods raise privacy concerns and may collect more information than necessary.
• Ethical and practical considerations necessitate optimizing data collection to the required amount.

Purpose of the Study:

• To determine the minimum duration of GPS data required to establish complete human activity spaces.
• To assess the point at which additional data yields negligible information gains.
• To introduce a novel metric for understanding measurement and information in human mobility.

Main Methods:

• Analysis of three weeks of GPS traces from over 100 student participants.
• Utilizing pairwise temporal analysis of Kullback-Leibler (KL) divergence of spatial histograms.
• Defining 'complete' activity space based on negligible marginal information gains.

Main Results:

• Fewer than 14 days of GPS data are sufficient to establish complete activity spaces.
• Convergence to low divergence levels occurs within days for all populations studied.
• Negligible information gain is observed in less than two weeks of data collection.

Conclusions:

• Optimizing GPS data collection duration is feasible and ethically important.
• The proposed metric provides valuable insights into data sufficiency for human mobility research.
• This research supports more efficient and privacy-conscious spatial behavior studies.