A spatiotemporal knowledge graph-based method for identifying individual activity locations from mobile phone data (2410.13912v1)

Abstract: In recent years, mobile phone data has been widely used for human mobility analytics. Identifying individual activity locations is the fundamental step for mobile phone data processing. Current methods typically aggregate spatially adjacent location records over multiple days to identify activity locations. However, only considering spatial relationships while overlooking temporal ones may lead to inaccurate activity location identification, and also affect activity pattern analysis. In this study, we propose a spatiotemporal knowledge graph-based (STKG) method for identifying activity locations from mobile phone data. An STKG is designed and constructed to describe individual mobility characteristics. The spatial and temporal relationships of individual stays are inferred and transformed into a spatiotemporal graph. The modularity-optimization community detection algorithm is applied to identify stays with dense spatiotemporal relationships, which are considering as activity locations. A case study in Shanghai was conducted to verify the performance of the proposed method. The results show that compared with two baseline methods, the STKG-based method can limit an additional 45% of activity locations with the longest daytime stay within a reasonable spatial range; In addition, the STKG-based method exhibit lower variance in the start and end times of activities across different days, performing approximately 10% to 20% better than the two baseline methods. Moreover, the STKG-based method effectively distinguishes between locations that are geographically close but exhibit different temporal patterns. These findings demonstrate the effectiveness of STKG-based method in enhancing both spatial precision and temporal consistency.

• The paper presents a novel spatiotemporal knowledge graph method that integrates spatial adjacency and temporal co-occurrence to enhance location identification.
• It uses a Fast Unfolding algorithm for community detection, improving spatial precision by identifying 45% more accurate activity locations.
• The study enhances temporal consistency by reducing start/end time variances by 10–20%, demonstrating potential for refined urban mobility analysis.

A Spatiotemporal Knowledge Graph-Based Method for Identifying Individual Activity Locations from Mobile Phone Data

Recent advancements in the analysis of human mobility have leveraged the growing availability of mobile phone data. Traditional approaches to identifying activity locations from this data have primarily relied on spatial clustering over prolonged timeframes. However, these methods often disregard the inherent temporal dynamics of individual movement, leading to inaccuracies in identifying distinct activity locations.

This paper introduces a novel approach utilizing a spatiotemporal knowledge graph (STKG) to more accurately identify activity locations by integrating both spatial and temporal dimensions. The STKG-based method leverages a structured graph model, which uses spatial adjacency and temporal co-occurrence relationships to map individual mobility. A community detection algorithm, optimized for modularity, identifies clusters of connected stays, representing potential activity locations.

Methodology and Implementation

The STKG method conceptualizes individual movements as nodes and edges in a knowledge graph, where spatial grids and temporal slots serve as foundational elements. By inferring relationships from mobile phone records, the spatial relationships within grids and temporal correlations between activity periods are established. The graph is subsequently partitioned using the Fast Unfolding algorithm to identify stays with dense connections, indicative of substantive activity locations.

A case paper conducted in Shanghai verifies the efficacy of this method. The results demonstrate critical improvements over baseline spatial and non-spatial methods. Notably, the STKG approach limits 45% more activity locations with the longest daytime stays within a reasonable radius, offering enhanced spatial precision. Temporal consistency is also improved, with start and end time variances reduced by 10-20% compared to traditional methods.

Implications

The proposed STKG-based method addresses the challenges of locational uncertainty and insufficient temporal consideration in mobile data analysis. By effectively distinguishing between proximal locations with divergent temporal patterns, the approach facilitates more nuanced activity pattern analyses. This method offers significant implications for urban mobility studies, enabling refined transportation planning and policy-making. As cities continue to evolve in complexity, the ability to accurately capture fine-grained mobility patterns becomes increasingly essential.

Future Directions

The research suggests further exploration into validation methodologies, including volunteer-based data comparison or aggregate-level validation against existing datasets. Moreover, the integration of STKGs into real-time urban planning initiatives could present substantial enhancements in predictive mobility models and personal travel behaviors.

By bridging the gap between spatial and temporal analytics, this approach sets a foundational paradigm shift for understanding and modeling human mobility through enriched data representations afforded by spatiotemporal knowledge graphs.