XSiteTraj: A cross-site user trajectory dataset.

Area of Science:

• Computer Science
• Data Science
• Social Network Analysis

Background:

• Social networking is integral to daily life, driving the need for effective user identification across platforms.
• Current user identification methods lack sufficient public datasets, hindering research and development.
• Existing datasets are often not publicly available, creating a significant bottleneck in the field.

Purpose of the Study:

• To introduce XSiteTraj v1.0, a novel, large-scale dataset for social network user linkage.
• To provide a valuable resource for advancing user identification algorithms and recommendation systems.
• To support research into user behavior patterns across different social media platforms.

Main Methods:

• The XSiteTraj v1.0 dataset was created by crawling data from Facebook, Foursquare, and Twitter.
• It includes over 27,000 users and more than one million check-in records globally.
• Each record contains timestamps, Points of Interest (PoIs), and geographical coordinates, with labeled identical users across sites.

Main Results:

• The dataset offers extensive spatio-temporal coverage, crucial for realistic user identification research.
• It enables the study of user behavior habits and the evaluation of cross-platform user identification algorithms.
• XSiteTraj v1.0 provides a benchmark for developing and testing new user linkage techniques.

Conclusions:

• The release of XSiteTraj v1.0 directly addresses the academic community's need for large-scale, public user linkage data.
• This dataset is expected to accelerate progress in social network user identification and related areas.
• Researchers can leverage XSiteTraj v1.0 for empirical studies on user behavior and algorithm performance evaluation.