Inferring human mobility using communication patterns.

Area of Science:

• Computational Social Science
• Network Science
• Urban Planning

Background:

• Understanding human mobility is vital for urban planning and disaster management.
• Existing mobility models often require accurate population data, limiting their applicability.
• Direct observation methods, like mobile phone tracking, raise privacy concerns.

Purpose of the Study:

• To investigate the inference of human mobility patterns solely from aggregated mobile phone call data.
• To develop and validate a model that predicts mobility without requiring explicit population counts.

Main Methods:

• Utilized aggregated mobile phone call data from Orange for Ivory Coast.
• Developed a predictive model based on call frequency between locations and geographical distance.
• Validated the model's ability to infer human mobility patterns.

Main Results:

• Human mobility patterns were accurately predicted by the proposed model.
• The model's performance was strongly correlated with call frequency and geographical distance.
• The model successfully inferred mobility using only aggregated call data.

Conclusions:

• Aggregated mobile phone call data can effectively predict human mobility patterns.
• The developed model offers a privacy-preserving alternative to traditional methods.
• Incorporating the social dimension of mobility through call data enhances predictive accuracy.