Measuring Public Life Through Digital Technologies: Investigating the Use of WiFi Sensing for Enhancing Public Space

다국어 초록 (Multilingual Abstract)

Vibrant public spaces are an indicator of the high quality of urban life. The success or failure of urban public spaces is dependent on how people use and interact with others in those spaces. Public life researchers have primarily measured the activities and behaviors of users through direct observation methods, reporting on findings that they have witnessed for themselves. Such simple techniques still have their place and are valuable but present their own set of multi-tasking challenges. For example, the researchers may be considering several themes; by nature, this approach is both labor-intensive and time-consuming. However, one digital technology, WiFi sensing, has recently been getting attention as a means for enhancing the research process by overcoming these data collection limitations. It passively detects the presence and location of a person who has a WiFi-enabled device 100 meters away around a WiFi sensor.

This thesis investigates the use of WiFi sensing as an alternative observation method for measuring different aspects of public life and public space. Many urban researchers have adopted sensing technology to analyze people’s mobility in various metropolitan areas. However, their findings are also relevant to quantifying people’s number and flow rate on the move. Currently, it seems unlikely that decision-makers will leverage this data to enhance the public space. This gap seems to arise from installing sensors and obtaining basic information without systematically considering how to apply the technology and the benefits to public life. These limitations align with the critical view of smart cities; sensing technology cannot transform the urban paradigm by offering new insights without moving beyond the restrictions of traditional approaches.

The feasibility of WiFi sensing is explored in a series of analyses regarding (1) the spatial and temporal properties representing people’s movements and behavior, (2) the possibilities of measuring alternative key metrics of public life, and (3) the practical applications for urban planning and design. Before presenting the findings, I proposed a conceptual framework that leverages the WiFi sensing and the public life survey framework, highlighting how the two fields are associated. I conducted several data-gathering experiments for the analyses, including WiFi and ground truth data using GPS. I then assessed the properties of WiFi data focusing on the spatial accuracy to positioning people’s location. Next, I tested stay point detection from WiFi traces and examined the accuracy of the WiFi stay points with GPS stay points as ground truth.

The thesis discusses the findings concerning the possibilities of WiFi sensing for public life studies in the context of a series of guided questions. First, based on our WiFi sensor network with an average sensor spacing of 50 meters, the WiFi data provides an approximate location of people within 20-30 meters accuracy at an interval of 30-second periods. WiFi sensing can also detect stationary activities, a critical metric in public life studies; However, the accuracy determining whether a person stays or moves from WiFi data did not achieve a high level of accuracy, with an F1 score of 0.384. The findings show that it is possible to locate people’s positions, at least at a street level, and then determine their behaviors, moving and staying patterns, with moderate accuracy. Several examples concerning the rhythm of public life and shopping-travel behavior were presented as practical case studies, helpful in urban planning and design.

In this thesis, I concluded that WiFi sensing provides an approximate location of people and their staying points with moderate accuracy. However, it effectively collects and analyzes long-term data at the neighborhood scale, supplementing the weaknesses mentioned above in manual observation. The analysis method proposed in this study improves the choice and range of research methods available for urban studies, expressly incorporating new sensing data and quantitative techniques into the standard toolkit of procedures used in public life studies. Specifically, the network of WiFi sensing with multiple sensors in a neighborhood can extend our understanding of public life beyond the named project area. This monitoring system also provides insights into how design elements, amenities, and programmed events enhance the vibrancy and vitality of public spaces by offering longitudinal evidence. This work contributes to transdisciplinary research on people’s mobility, which will in turn help to improve our overall understanding of cities, how they function, and how they are used in reality.

목차 (Table of Contents)

• Chapter 1. Research Background 1
• 1.1. Research background 1
• 1.2. Research questions 3
• 1.3. Research organization 5
• Chapter 2. Theoretical Background 7
• 2.1. Public life studies and emerging digital technologies 7
• 2.1.1. Public life studies 7
• 2.1.2. Public life studies through manual observation: strengths and weaknesses 11
• 2.1.3. Public life studies and emerging technologies: a review 13
• 2.2. WiFi sensing technologies in urban studies 16
• 2.2.1. Passive WiFi sensing 16
• 2.2.2. Literature review of urban studies based on WiFi sensing 17
• 2.2.3. A review of WiFi sensing as an observation method for studying public life 30
• 2.3. Conceptual framework: WiFi sensing for public life studies 31
• Chapter 3. Methodology 35
• 3.1. Data collection 35
• 3.1.1. Study areas 35
• 3.1.2. Experiment settings 39
• 3.1.3. Sensor 46
• 3.2. Data preprocessing 49
• 3.2.1. WiFi data 49
• 3.2.2. GPS data 52
• 3.3. Measurements 55
• 3.3.1. Spatial and temporal properties of WiFi data 55
• 3.3.2. Stay point detection 58
• Chapter 4. Results and Findings 62
• 4.1. Spatial and temporal properties of WiFi 62
• 4.1.1. The volume of data collected 62
• 4.1.2. Temporal properties of WiFi 63
• 4.1.3. Spatial properties of WiFi 65
• 4.1.4. Additional findings to spatial accuracy 69
• 4.2. Stay point detection evaluation 74
• Chapter 5. Application 77
• 5.1. Application list 77
• 5.2. Application 1: Rhythm of public life 79
• 5.2.1. Background 79
• 5.2.2. Data and method 79
• 5.2.3. Results and findings 80
• 5.2.4. Conclusion 84
• 5.3. Application 2: Shopping travel behavior 84
• 5.3.1. Background 84
• 5.3.2. Data and method 84
• 5.3.3. Results and findings 86
• 5.3.4. Conclusion 89
• Chapter 6. Summary and conclusions 90
• 6.1. Overview of thesis findings 90
• 6.2. Contributions of thesis 92
• 6.2.1. New insights on the understanding of public life and public space 92
• 6.2.2. Integrations of new digital data with standard data on urban studies 93
• 6.2.3. The contributions to leverage advanced technologies into urban studies 94
• 6.3. Future work and limitations 94
• Reference 103