Web-Centric Human Mobility Analytics

Introduction

Human mobility analytics has become increasingly important in various domains including urban planning, transportation, public health, and social science. With the rapid development of web technologies and the emergence of large language models (LLMs), web-centric approaches to human mobility analytics have gained significant attention. This tutorial provides a comprehensive overview of web-centric human mobility analytics, covering methods, applications, and future directions in the LLM era. We discuss how web data sources, web-based analytical techniques, and web platforms can be leveraged to understand, model, and predict human mobility patterns. Furthermore, we explore how LLMs are transforming this field by enabling more sophisticated natural language understanding and generation capabilities for mobility data analysis.

Intended Audience

This tutorial is designed for professionals, researchers, and practitioners interested in human mobility analytics, assuming basic knowledge of AI and data analysis.

Attendees will gain understanding of mobility analytics methodologies across location, individual, and macro levels, plus advanced frameworks like federated learning and causal inference for real-world applications.

Tutorial Outline

S1. Welcome and Introduction (10 mins)

1.1 Overview of Human Mobility Analytics

S2. Location-Level Human Mobility Analytics: (40 mins)

2.1 Predictive Modeling for Next Location Visits
2.2 Applications of Location-Level Analytics

S3. Individual-Level Human Mobility Analytics: (40 mins)

3.1 Trajectory Compression
3.2 Trajectory Recovery

S4. Macro-Level Human Mobility Analytics: (40 mins)

4.1 Regional Crowd Flows Modeling
4.2 Universal Knowledge Modeling

S5. Advanced Learning Framework: (30 mins)

5.1 Federated Human Mobility Learning
5.2 Spatio-Temporal Continuous Learning
5.3 Foundation Models (e.g., LLMs) for Mobility: Challenges, Applications, and Directions

S6. Conclusion and Open Discussion: (20 mins)