Handbook of Mobility Data Mining, Volume 1: Data Preprocessing and Visualization

Handbook of Mobility Data Mining, Volume One: Data Preprocessing and Visualization introduces the fundamental technologies of mobile big data mining (MDM), advanced AI methods, and upper-level applications, helping readers comprehensively understand MDM with a bottom-up approach. The book explains how to preprocess mobile big data, visualize urban mobility, simulate and predict human travel behavior, and assess urban mobility characteristics and their matching performance as conditions and constraints in transport, emergency management, and sustainability development systems. The book contains crucial information for researchers, engineers, operators, administrators, and policymakers seeking greater understanding of current technologies' infra-knowledge structure and limitations.

Further, the book introduces how to design MDM platforms that adapt to the evolving mobility environment, new types of transportation, and users based on an integrated solution that utilizes sensing and communication capabilities to tackle significant challenges faced by the MDM field. This volume focuses on how to efficiently pre-process mobile big data to extract and utilize critical feature information of high-dimensional city people flow. The book first provides a conceptual theory and framework, then discusses data sources, trajectory map-matching, noise filtering, trajectory data segmentation, data quality assessment, and more, concluding with a chapter on privacy protection in mobile big data mining.

• Introduces the characteristics of different mobility data sources, like GPS, CDR, and sensor-based mobility data
• Summarizes existing visualization technologies of the current transportation system into a multi-view frame, covering the perspective of the three leading actors
• Provides recommendations for practical open-source tools and libraries for system visualization
• Stems from the editor’s strong network of global transport authorities and transport companies, providing a solid knowledge structure and data foundation as well as geographical and stakeholder coverage

• Language: English
• Publisher: Elsevier Science
• Release date: Jan 29, 2023
• ISBN: 9780443184291

Book preview

Chapter One: An overview of urban data variety and respective value to urban computing

Hang Yin     LocationMind Inc., Chiyoda-ku, Tokyo, Japan

Abstract

Urban data provide sufficient information for urban planning, transportation management, urban anomaly analysis, etc., which can help benefit the lives of residents. Recently, various urban data have been widely collected and analyzed by machine learning algorithms, through which further information and conclusions are generated for guiding urban management. This process is called urban computing. In this survey, we make a summary of various types of urban datasets obtained from diverse devices, that is, trajectory, trip records, call detail record, urban sensor record, event record, environment data, social media, and surveillance camera data, which are also followed by value analysis in the field of urban computing.

Keywords

Data format; Data source; Urban computing; Urban management

1. Introduction

The rapid progress of urbanization has led to many big cities, which benefits the lives of residents but also causes some big challenges, such as traffic congestion, environmental pollution, and energy consumption. Nowadays, based on sensing technologies and large-scale computing infrastructures, we can get various urban data from different sources, which implies rich knowledge about a city and can help solve these challenges. Meanwhile, urban anomalies concerning traffic, environment, and crowds may result in loss of life or property if not handled properly. Recently, urban data-driven anomaly analysis frameworks have been forming, which utilize urban big data and machine learning algorithms to detect and predict urban anomalies, thus reducing the losses or mitigating impact. These processes based on urban data collection and computer analysis belongs to the field of urban computing [3].

1.1. Definition of big data

Big data is a buzzword that permeates much of the discussion surrounding the use of digital information in business and government. However, there is almost no consensus on the concept definition of the term. Therefore, we provide some alternative views about what constitutes big data.

• Volume: Big data is often characterized by its huge size, often in petabytes or more. Urban data collected from smartphones is a typical example of a huge size in the concept of time, space, and quantity.

• Velocity: Big data is often generated continuously, in or near real time.

• Variety: Big data can be of any type, so its analysis methodologies need to be flexible enough to deal with any type or, increasingly, a mix of these.

1.2. Definition of urban data

Nowadays, smart devices and various kinds of sensors widely located in cities collect the data produced in urban areas, satisfying the basic views of big data but also being of some unique qualities [1].

• Huge volume: Many urban activities leave a rich digital footprint, such as vehicle tracks and posts on social media platforms, which contain a large amount of urban data.

• Various forms: Different sources usually produce different forms of urban data, including structured data such as human trip records and unstructured data such as surveillance videos.

• Unique dimension: Urban data are associated with time stamps and location labels: Therefore, urban data often contains rich contextual information and brings complex spatio-temporal associations and dependencies between different data points [4].

1.3. Urban computing and challenges

As Fig. 1.1 shows, urban computing is the process of acquiring, integrating, and analyzing urban data generated from different sources (such as sensors, devices, vehicles, buildings, and humans) in urban range to solve major problems concerning about city (such as traffic congestion, air pollution, crowd gathering) [5]. The goal of urban computing is to improve the urban environment, human life quality, and city operation systems. Urban computing is an interdisciplinary field fusing the computing science field with traditional fields, including transportation, environment, economy, etc. [6], which can help us not only in the city management field but also understand urban phenomenon and improve the concept of the city development [7].

However, the goal of urban computing and its data-driven methodologies result in the following challenges:

• Data acquisition technology can continuously collect data unobtrusive across the city. It's easy to monitor traffic flow in sections, but it's a challenge to continuously detect citywide traffic because we don't have sensors on every section. This could be achieved by building a larger sensor infrastructure but would, in turn, increase the burden on the city, including the energy consumption, sensor expense, and post of the observer.

Figure 1.1  Relation among Urban Planning, Urban Computing, Urban Data, and Urban Area. Credit: original.

• For data collection methods that require user participation, privacy protection is an issue that cannot be ignored. In the perception process, users contribute their personal data (usually from smartphones), which is collected by the system to the cloud for analysis. In this process, an necessary balance should be kept among energy consumption, privacy protection, and the efficiency of sharing data.

• The data generated by traditional sensors are generally well structured, easy to understand, and of relatively little missing data. However, the data provided by users usually have great uncertainty and noise, such as data formats with great freedom like text and image.

2. Urban data variety and value

Urban data is the basis of all due to the data-driven methodologies of the urban computing field. Nowadays, various data can be collected from electronic devices. One of the most important sources is smartphones. For example, when a user makes a phone call or connects to a cellular network, the location will be recorded in real time and can be used to track the user's mobility. The information posted by smartphones on social media is also one of the sources of information that can be accepted or analyzed by the system. In addition, electronic sensors distributed around urban areas are also an important source of high-quality urban data. For example, surveillance cameras can record and store the surrounding scene, and traffic sensors can provide real-time traffic conditions. In this section, we divide urban data sets into eight categories based on data types and attributes. Six structured urban data contains trajectory, human trip record, CDR (mobile phone call detail record), event record, urban sensor record and environmental monitoring data, and two unstructured data contains social media data and surveillance camera data.

2.1. Trajectory

The trajectory is a track consisting of a series of time and position records that are recorded and reported by a GPS device. The trajectory data set provides the most detailed and comprehensive record of object movements [8]. A trajectory:

represents a trip of a moving object with a sequence of location and time pairs. and are the origin location and start time of the trip, while and are the destination location and arrival time of the trip. The location of the rest of the points in the trajectory are called via locations of . The means a range of urban trajectory data correction, all locations of , and should be contained in this range [9].

Vehicles equipped with GPS devices provide an important type of movement, given that public and private vehicles are the main transportation means for urban human mobility [10]. People use vehicles for commuting to and from work, for regular and irregular chores, and for leisure activities. By analysis of the observed movements, researchers strive to better understand the demographics of a city, the distribution of services around a city, the effectiveness of the various transportation networks, the dynamics of traffic conditions, the different driving behaviors, etc [11,12].

Take trajectory data collected by taxi with GPS devices as an example. As one of the most important of urban transportation, taxi is ubiquitous in urban areas and operates almost 24hours a day. The taxi service equipped with a GPS device is driven by the diverse needs of a large number of people, and the movement trajectory of the taxi is not limited by the fixed-route [13,14]. A taxi's GPS trajectory can tell us quite precisely where a passenger got on, off, which route to take, and what steps a driver took to find a new passenger before picking up a passenger. With the diverse needs of passengers and the nature of the service in continuous operation, taxi GPS tracking provides rich and detailed data for the analysis of the motivation and behavior of the urban floating population