[Paper Review] A New Insight into Land Use Classification Based on Aggregated Mobile Phone Data

Tao Pei, Stanislav Sobolevsky+3 more|arXiv (Cornell University)|Oct 23, 2013

Human Mobility and Location-Based Analysis7 references4 citations

TL;DR

This paper proposes a novel land use classification method using aggregated mobile phone data, leveraging hourly call volume patterns and relative activity rhythms to infer urban land use types via semi-supervised fuzzy c-means clustering. Evaluated in Singapore, the approach achieved a 58.03% detection rate, demonstrating that mobile phone data can effectively capture social function-based land use distinctions, especially in areas with higher tower density.

ABSTRACT

Land use classification is essential for urban planning. Urban land use types can be differentiated either by their physical characteristics (such as reflectivity and texture) or social functions. Remote sensing techniques have been recognized as a vital method for urban land use classification because of their ability to capture the physical characteristics of land use. Although significant progress has been achieved in remote sensing methods designed for urban land use classification, most techniques focus on physical characteristics, whereas knowledge of social functions is not adequately used. Owing to the wide usage of mobile phones, the activities of residents, which can be retrieved from the mobile phone data, can be determined in order to indicate the social function of land use. This could bring about the opportunity to derive land use information from mobile phone data. To verify the application of this new data source to urban land use classification, we first construct a time series of aggregated mobile phone data to characterize land use types. This time series is composed of two aspects: the hourly relative pattern, and the total call volume. A semi-supervised fuzzy c-means clustering approach is then applied to infer the land use types. The method is validated using mobile phone data collected in Singapore. Land use is determined with a detection rate of 58.03%. An analysis of the land use classification results shows that the accuracy decreases as the heterogeneity of land use increases, and increases as the density of cell phone towers increases.

Motivation & Objective

• To explore the potential of aggregated mobile phone data as a complementary data source for urban land use classification.
• To address the limitation of remote sensing methods that focus only on physical characteristics by incorporating social function indicators from human mobility.
• To develop and validate a clustering-based method that uses temporal patterns of mobile phone activity to infer land use types.
• To assess the influence of land use heterogeneity and cell tower density on classification accuracy.

Proposed method

• Construct a time series of aggregated mobile phone data, capturing hourly relative call patterns and total call volume per geographic unit.
• Apply a semi-supervised fuzzy c-means clustering algorithm to group spatial units based on their temporal activity profiles.
• Use the temporal dynamics of mobile phone usage—such as peak hours and daily rhythms—as proxies for social function of land use.
• Validate the clustering results against ground-truth land use data from Singapore.
• Incorporate spatial constraints through semi-supervision to improve cluster interpretability and accuracy.
• Define cluster centroids based on activity patterns, enabling identification of residential, commercial, and mixed-use zones.

Experimental results

Research questions

• RQ1 Can aggregated mobile phone data effectively represent the social function of urban land use types?
• RQ2 How does the accuracy of land use classification vary with the heterogeneity of land use in a given area?
• RQ3 To what extent does the density of cell phone towers influence classification performance?
• RQ4 Can temporal patterns of mobile phone activity reliably distinguish between different land use types?
• RQ5 How does the proposed method compare to traditional remote sensing approaches in capturing functional land use?

Key findings

• The proposed method achieved a land use classification detection rate of 58.03% using mobile phone data from Singapore.
• Classification accuracy decreased as land use heterogeneity increased, indicating challenges in distinguishing mixed-use areas.
• Higher cell tower density was associated with improved classification accuracy, suggesting better data resolution enhances performance.
• Temporal patterns of mobile phone activity, such as peak hours and daily rhythms, provided strong indicators of land use function.
• The integration of social function data from mobile phones complements physical characteristics captured by remote sensing, offering a more holistic classification approach.
• The semi-supervised fuzzy c-means clustering effectively grouped spatial units based on behavioral patterns, enabling interpretable land use inference.