People have observed and investigated dynamic behaviours of movement through the ages. In today’s world, movement data have been more collectively generated particularly due to the development and the deployment of location-aware technologies and geo-sensor networks, the emergence of ubiquitous computing environments, and the usefulness of these technologies in everyday life. This book addresses data mining challenges of extracting valid, understandable and useful knowledge from large-scale data about movement. Such knowledge can facilitate decision-making processes in various application fields including urban planning, transportation, animal ecology, logistics and location-based services.
This book is about movement data analysis specifically focusing on visual analytics that leverages computational processing and human intelligence to transform large and complex movement data to information and knowledge. The authors well present the topic with 9 chapters that cover a general overview (Chapter 1), a conceptual framework (Chapter 2), data preprocessing (Chapter 3), basic visualization techniques (Chapter 4), a variety of visual analytic methodologies (Chapters 5–8) and concluding remarks (Chapter 9).
Chapter 1 illustrates the main concept of the book. This chapter consists of various illustrations, which are derived from results of several visual analytics using GPS data of single vehicle and multiple vehicles. The concrete examples and illustrations can make readers easy to grasp the main concept and the capability of visual analytics that extracts semantic-rich information from typically semantic-poor raw movement data.
Chapter 2 discusses a conceptual framework and provides the fundamental and comprehensive understanding of complex movement data. The chapter describes the framework by types, characteristics and relationships related to movement data, types of movement behaviours and types of analytical tasks. The chapter starts by introducing fundamental sets, space, time and objects, and then defines two basic types of spatio-temporal objects, moving objects (movers) and spatial events. Based on the characteristics of fundamental sets and types of objects, the authors introduce four perspectives of the phenomena of movement, mover perspective, space perspective, time perspective and spatial event perspective.
Chapter 3 focuses on data preprocessing, which transforms raw data to usable forms. Techniques covered in this chapter include interpolation, re-sampling, division of trajectories, alignment of temporal and spatial references, derivation of new thematic attributes, extraction of movement events, generalization, simplification and aggregation. These are necessary data pre-processes for further analyses including but not limited to visual analytics.
Chapters 4–8 present methodological and technical aspects. Chapter 4 introduces basic visualization techniques based on cartography, space–time cube, time graph and temporal bar chart. The chapter also describes filtering techniques that reduces cluttering and occlusion effects and highlights specific information of interest. Chapters 5–8 focus on visual analytic methodologies and techniques based on four different perspectives, mover perspective, spatial event perspective, space perspective and time perspective. This ‘multi-perspective view of movement’ is the key idea of the book.
The authors conclude the book with Chapter 9 that discusses the link between the conceptual framework and methodologies, privacy concerns and future perspectives. The summary of correspondence between task foci and presented methods for data transformation, analysis and visualization re-highlights the key idea of this book, multi-perspective view of movement. Regarding privacy concerns, the authors outline approaches to protect privacy information depending on the types of analysis tasks. Moreover, the authors point out the current need for visual analytic methods to reconstruct routine movement behaviours of people by transforming their trajectories from geographical space to abstract space. Finally, the authors provide the summary of open problems and directions for future research that helps researchers who are seeking state-of-the-art research agendas in the visual analytics of movement.
This book is thoughtfully conceived, systematically organized and written in easy-to-read style with numerous illustrations and useful reference materials. The book includes conceptualization, modelling, transformation and preprocessing of movement data, which can be further applied to not only visual analytics but also other data mining exercises. The authors present state-of-the-art trajectory data mining techniques including pre-clustering procedure, scalable clustering, spatial or spatio-temporal clustering and progressive clustering, and novel visualization techniques such as flower diagram and growth ring maps together.
The value of this book is in not only presenting technical and methodological aspects of visual analytics but also profound discussions on the applicability of methods and interpretations of analytical results. A variety of movement data is used in the book including single and multiple cars, vessels, public transportations, visitors in a car race event, wild animals, laboratory mice and geo-tagged flicker and twitter data. These data were collected by four different methods: change-based, time-based, location-based and event-based. The authors discuss the applicability of methods to various types of data-sets. For each method, the authors provide detail interpretations of analytical results together with illustrations and the pros and cons.
Software developers can find technical and methodological details through texts, pseudo codes and diagrams of data types and relationships, which will certainly help developing algorithms, methods and applications. This book does not cover technical details such as computational setting (e.g., hardware), programming environment, database structure and query performance or design and structure of software interface. In addition, high performance computing (HPC) solutions, in the forms of distributed and/or parallel computing, are out of scope of this book.
With systematic presentations, various examples and many illustrations, practitioners such as data scientists can easily grasp ideas of how to extract and present useful and understandable knowledge from movement data. The profound discussions are extremely useful to identify appropriate methodologies for their data and their needs.
Furthermore, this book can also be recommended to a wide range of readers who are interested in not just movement data but general spatio-temporal phenomena or general data mining techniques since presented techniques can be applicable to general spatio-temporal or even more complex data-set.
Overall, this book is an excellent contribution to data mining and GIS communities. It should appeal to not only research scholars who are specialized in movement data analysis but also software developers, practitioners and general audiences who are dealing with movement data or just interested in movement behaviours.
Atsushi Nara
Department of Geography, San Diego State University, San Diego, CA, USA