March 16th, 2007 by Coral Gehrke for GEO 565: Geographic Information Systems and Science, Oregon State University
My graduate research project involves a visitor tracking study at the Oregon Coast Aquarium, a free-choice learning (FCL) setting. According to the National Science Foundation, free choice learning is “the life-long process in which every person acquires knowledge, skills, attitudes, and values from daily experiences and resources in his or her environment…[It] is self-directed, voluntary, and motivated mail by intrinsic interest, curiosity, exploration, and social interaction.” This type of learning occurs outside of school, encompassing approximately 97% of a person’s lifetime. Aquaria, zoos, science centers, museums, national parks, marine parks, magazines, television programs, newscasts, traveling exhibits and internet sites, among many others, are all examples of FCL settings or opportunities. Part of my data analysis examines how visitors use the space of the aquarium, particularly how they move through each gallery and through the aquarium as a whole. In working on this analysis, I have become increasingly interested in how GIS might be used in FCL research, particularly in performing spatial analyses to help identify and understand visitor usage patterns. Below you will find the results of a preliminary literature review with articles focusing either on how GIS is currently being utilized in FCL environments or on how GIS is being used to track and analyze peoples’ movements and usage patterns in other settings. I am hopeful that these might provide a framework for this potential application of GIS in mapping and analyzing visitor usage patterns in and navigation through free-choice learning environments.
Abstract summary: New, highly accurate location determination technologies that can be used indoors as well as outdoors will allow for new spatial applications and analysis that track people in new and different application areas in real time. I think this assertion, while limited in detail, is an important one to note. Many of the articles below address many of the limitations of tracking technology, particularly with respect to spatial accuracy and the ability to function indoors. It is therefore quite relevant to the future uses and applications of GIS in indoor free-choice learning settings such as museums, science centers, etc.
Shoval, N. and M. Isaacson. 2006. Application of Tracking Technologies to the Study of Pedestrian Spatial Behavior. The Professional Geographer 58(2): 172–183.
This article looks at the potential use of satellite navigation systems and land-based navigation systems, two of the main tracking technologies currently in use, in gathering data on pedestrian spatial behavior as an alternative to traditional methods involving time-space diaries kept by the subjects themselves. Recent advances in GIS have resulted in an increase in the number of studies analyzing “place and space,” particularly in the social sciences. The authors state that the time-space budget technique procedures in use today are methodologically flawed because they require the subject actively record his or her movements during the experiment, interrupting “normal” behavior. Additionally, reliability and quality of data varies depending on the diligence of the subject and it can often be difficult to find participants for such studies. The authors discuss reasons for the limited use of advanced technologies in gathering data on human spatial behavior and how the advent of smaller, passive technologies might be changing this. They go on to discuss several advantages, disadvantages, and limitations of both GPS and land-based tracking systems, as well as the use of hybrid solutions. In their experiment they tested and compared the performance of two of these technologies, specifically GPS and TDOA (time difference of arrival) land-based antenna system in Tel Aviv-Jaffa, Israel. The study found the GPS is more accurate than the TDOA, and is ideal for use in studies performed on smaller geographic scales and/or with high data resolution needs. However, TDOA, when compared to GPS, is small, lightweight and does not require much attention from the subject and is ideal for use in situations that do not require more accuracy than that provided by the GPS. The authors conclude that both technologies are “within easy reach and could be readily exploited to study pedestrian spatial behavior.”
Kwan, M. 2000. Analysis of human spatial behavior in a GIS environment: Recent developments and future prospects. Journal of Geographical Systems 2:85-90.
In this paper, the author looks at how developments and increased availability in metropolitan and individual level geographic data sets as well as GIS based spatial analysis tools provide new opportunities for social scientists in analyzing and understanding human spatial behavior. The author discusses spatial analysis of human behavior in the past, recent analytical developments, new analytical opportunities, as well as future prospects in the analysis of human spatial behavior. In the past, methods of studying human spatial behavior have been limited by the data and tools available, leading to unrealistic or overly simplistic models and theories with respect to behavioral complexities and the complexity of real urban environments. Additionally, the spatial frameworks utilized in past studies could not be used for analysis at the disaggregate and individual levels and typically did not incorporate data about a person’s actual spatial knowledge of his or her environment. These difficulties may be less of problem in the future as data about urban environments, individual-level geo-referenced data, new data collection methods, and analytical tools continue to become more readily available. The author states that these changes allow for more realistic representations of the complex objective environments in a GIS environment as well as allowing for more subjective factors affecting human spatial behavior to be incorporated into the GIS. Researchers will be able to discover the rules or principles underlying these behaviors. Finally, the ability to focus on the individual will allow researchers to look at fine-scale, inter-personal differences in “socially significant’” categories.
Kwan, M. and J. Lee. 2003. Geovisualization of human activity patterns using 3D GIS: A Time-Geographic Approach. In M.F. Goodchild and D.G. Janelle. Eds. Spatially Integrated Social Science: Examples in Best Practice, Chapter 3. Oxford: Oxford University Press.
Time-geography has been used by social scientists to analyze human activity patterns and movements in space and time in a continuous temporal sequence in geographical space. Researchers can also use time-geography analytical methods to study interactions between space and time and how they influence human activity patterns in a particular location, however few studies have taken advantage of this methodology because of lack of data on the individual level and the difficulties associated with accurately representing complex environments. Improvements in GIS and the increased availability of data (see previous annotation for more details) have resulting in a resurgence and further development of time-geographic methods. The authors discuss the concepts of scientific visualization and geovisualization. Using GIS-based 3-D geovisualization methods in analyzing human activity patterns prevents the loss of important data that occurs with 2D maps, particularly things like information about timing, duration, and sequence of activities or trips. Additionally, GIS is flexible, has the capability of performing spatial analyses, can integrate large amounts of data from different sources while generating complex and accurate (realistic) representations of actual environments, and retains the complexity of the original data. The authors go on to describe a study on gender/ethnic differences in space-time activity patterns in Portland, Oregon using a time-geography methodology. The study shows that geovisualization methods can help researchers understand interactions that occur between space and time and human spatial behavior as well as other social characteristics/demographics. They conclude with some of the difficulties encountered in using this methodology.
Zhao, H. and R. Shibasaki. 2005. A novel system for tracking pedestrians using multiple single-row laser-range scanners. IEEE Transactions on Systems, Man, and Cybernetics – Part A: Systems and Humans. 35 (2).
Video recordings are currently used to collect crowd data in a variety of settings; however these methods are limited by the number, placement, orientation, resolution, viewing angles, settings, changes in illumination, calibration, etc. of the cameras. The authors of this paper propose using the newly developed technology of small single-row laser-range scanners (LD-A) to track pedestrian movements in wide, open areas (i.e. railway stations, shopping malls, exhibition halls, etc.) and analyze changes in visitor flow patterns throughout the day. The lasers used in this study were placed on the floor and scanned still objects and moving feet approximately 20cm above the ground. The authors developed a tracking algorithm and used a pedestrian’s walking model, a definition of state model, and a filter tracing process to analyze the data collected by the LD-A. This data was then spatially and temporally integrated into a global coordinate system. The authors used real experiments and computer simulations to evaluate the new system. They found that while the algorithm is not robust or accurate enough to track each individual and complete crowd trajectories, they system can be used to track and analyze over-all crowd flow. Advantages of this technology include that it is a direct measurement, the system is easily calibrated and converted into a real coordinate system, it is low cost, will likely have the ability to function in real-time in the near future, and it avoids privacy issues that can arise when video images are used.
In this pilot tracking study, GPS and GIS technologies were used to gather and display data on spatial properties (including position, speed, distance, time, & elevation) and behavior of a civilian over a two-week period in what the authors call location-based intelligence. This type of data can currently be gathered by location based service (LBS) applications providers, particularly those related to telecommunications. The authors describe different methods of tracking people including GPS, Wi-Fi, video monitoring, etc. as well as how this information can be stored and displayed (spatially represented) in a GIS in the section on location-based surveillance. With an increasing number of private companies investing in these types of technologies, the authors point out that the context of how they will be used, for what purposes, and liability and privacy issues must be considered. Near the end of the article, the authors point out the different types of detailed personal information that can be constructed using the data collected in this study, including residence, work and social activities, as well types and times of activities engaged in. Some of the common technical issues as well as ethical issues that can arise when using GPS tracking include accuracy, the need/ability to edit tracks in the GIS, the creation of intelligent systems that “know” when behavior is unusual, level of detail of supporting datasets (i.e. roads, buildings, etc.), and user awareness.
The authors of this paper suggest the development of a system that provides personalized recommendations, directions, etc. for each visitor to a museum or archaeological site by creating a network with the ability to predict user behavior using topological information from a GIS to create a more stimulating, accessible, and rich experience based on the individual visitor’s preferences. The authors argue that GIS supplies the necessary technological support to manage, transform, retrieve, and analyze the large amounts of data needed to make such a system work. A diagram of the system architecture is included in the article and is helpful in understanding how the system would be engineered. As part of this system, a visitor may use the Museum Wearable which is an audiovisual device that provides a narration for the exhibit that is interactive in time and space. The data collected by the positioning system that tracks visitor movement and location would be stored in the Usage Log and used to identify visitor usage patterns. These patterns would provide the basis for the Recommendation System and the Prediction Subsystem, which would provide the personalized visitor recommendation. Additionally, the log would provide invaluable information for exhibit and institutional evaluation projects. They also propose a plan for the integration a solution to the one of the main limitations of tracking using GPS that would allow tracking and therefore personalization of visits indoors as well as out.
O’Connor A., A. Zerger, and B. Itami. 2005. Geo-temporal tracking and analysis of tourist movement. Mathematics and Computers in Simulation 69: 135–150.
This study is of particular interest to me because much of their analysis parallels what I am doing and attempting to do with in my own analysis (total time spent, order in which gallery components are visited, time spent at components, “typical” visit/flow patterns).
Many recreational resource managers are using agent-based simulators, such as the program recreation behavior simulator (RBSim), to develop models of tourist movement patterns to analyze current uses, predict future demands, and manage impacts in parks and other outdoor recreational settings. Tracking actual visitors is difficult because of the potentially large number of people within a given area, the variety of choices available to the visitor as to what to do, where to go, etc., and the difficulty associated with tracking movement without affecting “normal” behavior. The authors of this study used the Alge timing system (human tracking technology used in running events and triathlons) to track visitors at the Twelve Apostles, Port Campbell National Park, Victoria, Australia, in conjunction with GIS and statistical analysis techniques to identify tourist typologies (defined simply as a common characteristic) to be used in RBSim. They analyzed the data for travel time, total time spent, estimates of time spent at a specific location, travel sequence, and the potential influences of crowding on behavior. The authors found that the Alge technology works well in collecting data in outdoor recreation areas and make suggestions on how their data collection methods and analysis might be improved.
In this analysis (please see above annotation for more details), the authors were successful in identifying typologies with respect to the sequences of travel. They initially identified five mutually exclusive, movement categories or “typical trips” and then compared these to the actual paths taken by visitors. Visitors followed each of the path types; however, the majority of visitors followed the Type 1 trip. These same patterns and proportions were observed on multiple days. The results indicate that tourist behavior, in some cases, can be divided into distinguishable groups based on movement sequences. Visualizations of visitor path sequences (spatial patterns of movement) support this conclusion. Finally, the authors point out the usefulness of and suitability of GIS related technologies in addressing park management questions.
This master’s thesis demonstrates that GIS and related technologies can be used to provide good estimates of trail usage patterns in national parks and other outdoor recreation areas. Information on the amount, type and distribution of day use for two trails in Yellowstone National park was collected with mechanical counters, visual observations and GPS units. Data on profiles of the hikers, distance traveled, hiking/trail behavior was analyzed using GIS. The analysis showed that differences in hiker profiles, with respect to age, speed, party size, and activity, were dependent upon the distance covered, as was trail behavior (leaving the trail). Not surprisingly, the profiles of hikers that stayed and those that left the trail varied according to group size and activity. The thesis provides further support that GPS and GIS technologies are useful and informative tools for park and outdoor recreation managers.
The authors demonstrate the use of a GIS as a tool in storing and processing visitor use data to support management in the national parks in Japan, particularly with respect to erosion issues. Visitor use and congestion was tracked using GPS and with gate counts. The data collected was entered into the Mt. Sibutsu Management System (SMS) developed using ArcGIS. The authors successfully used ArcGIS Tracking Analyst to show actual visitor use and to identify the timing and areas of congestion (which can lead to soil erosion in some areas of the park). Managers can use the SMS to develop land protection strategies focused on the areas identified as being at risk. Additionally, Tracking Analyst can be used to simulate trail closures, levels of congestion, etc. The article shows the data model used in the analysis.