Big Data and Mobility as a Service FIRST EDITION Haoran Zhang Center for Spatial Information Science, The University of Tokyo, Kashiwa, Chiba, Japan Xuan Song Department of Computer Science and Engineering, South University of Science and Technology, Nanshan, China Center for Spatial Information Science, The University of Tokyo, Kashiwa, Chiba, Japan Ryosuke Shibasaki Center for Spatial Information Science, The University of Tokyo, Kashiwa, Chiba, Japan Table of Contents Cover image Title page Copyright Contributors Introduction 1: Background 2: Big data: Definition, history, today 3: MaaS: Definition, history, today 4: Big data X MaaS 5: Summary Chapter 1: MaaS system development and APPs Abstract 1: The development history of MaaS 2: The category of MaaS system 3: Study case 4: Future development trend of MaaS system Chapter 2: Spatio-temporal data preprocessing technologies Abstract 1: Introduction 2: Raw GPS data and workflow of data preprocessing 3: Key technologies and corresponding application 4: Case study 5: Conclusion Chapter 3: Travel similarity estimation and clustering Abstract 1: Introduction 2: Trajectory similarity 3: Travel pattern similarity 4: Origin-destination matrix similarity 5: Case study 6: Conclusion and future directions Chapter 4: Data fusion technologies for MaaS Abstract Acknowledgments 1: Introduction 2: Data formula 3: Categories of data fusion methods in MaaS 4: Data fusion based on deep learning 5: Decomposition-based methods 6: Challenging problems of data fusion in MaaS 7: Conclusions Chapter 5: Data-driven optimization technologies for MaaS Abstract 1: Overview of data-driven optimization for the urban mobility system 2: Overview of the general concept in MaaS System 3: Mobility resource allocation in MaaS system 4: Data-driven optimization technologies for resource allocation in MaaS 5: Real-world application and case study 6: Conclusions Chapter 6: Data-driven estimation for urban travel shareability Abstract Acknowledgment 1: Introduction 2: Emerging sharing transportation mode 3: Background to traditional data and their limitations 4: New and emerging source of data 5: Emerging form of key technologies 6: Case study of ABM in urban shareability estimation 7: Opportunities and challenges 8: Conclusions Chapter 7: Data mining technologies for Mobility-as-a-Service (MaaS) Abstract 1: Introduction of data mining technologies in MaaS system 2: Data mining technologies in MaaS system 3: Methodologies of data mining technologies used in MaaS system 4: Case study of data mining for MaaS: Bike sharing in Beijing during Covid-19 pandemic 5: Summary of chapter Chapter 8: MaaS and IoT: Concepts, methodologies, and applications Abstract 1: Introduction 2: Overview of the concept 3: Key technologies and methodologies 4: Application and case study 5: Conclusion and future directions Chapter 9: MaaS system visualization Abstract 1: Overview of the general concept 2: The key visualization technologies in MaaS for different stakeholders 3: Real-world application and case study 4: Conclusion and future directions Chapter 10: MaaS for sustainable urban development Abstract 1: Introduction 2: MaaS interacted with urban traffic and space 3: Strategies for MaaS in urban sustainable development at multiple scales edge computing 235–236 future research directions 240–241 intelligent transportation equipment 231–232 microservices 232–234 overview 230–231 security technologies 236–237, 237t case 7–21 Moovit 15–17 Uber 17–21 UbiGo 10–12 Whim 12–15 category 5–7 booking and payment, integration of information integration no integration service offering, integration of 6–7 societal goals, integration of data-driven optimization technologies for characteristics of 150t customized bus (CB) application 164–170 ecosystem 151t mathematical notations 155t mobility resource allocation in 152–157 overview of 149–152 resource allocation in 157–163 urban mobility system 143–149 data mining technologies for See Data mining technologies, for mobilityas-a-service (MaaS) development history alliance 2–3 conception development early application 1–2 innovation 4–5 revolution 4–5 future development trend 21–23 data-integrated 21–22 future-oriented 22–23 sustainable 23 heterogeneous multisource data 245–246, 246t overview 245–247 real-world application and case study 254–261 city manager 257–258, 257f demanders of mobility 254–255, 254f open-source visualization tools and libraries 258–261 transportation service supplier 255–256, 256f urban spaces of 266–267, 268f visualization technologies 247–254 city manager 252–254 demanders of mobility 247–248, 247t, 249f transportation service suppliers 248–252, 250f Mobility pattern 90 Mobility resource allocation, in MaaS system 152–157, 153–155f Mobility services 151 Mobility tableau 94 Mobmap 258–259, 258f Mobsimilarity 98–103, 99f, 102f Monte Carlo method 144–145 Moovit characteristics big data technologies 17 integration 16 price 16 sustainability 17 overview 15 services bike routes 16 digital payment 15 live navigation 16 real-time alerts 16 real-time arrivals times 15 stops’ visualization 16 user reports 16 Movement data 21 Multiagent reinforcement learning (MARL) 190 Multiagent systems (MAS) 187–188 Multicentercenter cities 271 Multiple-mode transportation data visualization 261 Multisource geospatial big data 183–184, 184f N Navigation data 186 Newsvendor problem (NVP) 161–162 New York City (NYC) 145–146 Nonlinear support vector machine 212–213, 212f Nonnegative matrix factorization (NMF) 93 Normalized mass angle (NMA) 102 Numerical variable 221 O One way distance (OWD) 86 Open-source visualization tools and libraries 258–261 Operation data 208–209 Origin-destination (OD) 246, 259, 259f matrix similarity 93–103, 94f tableau similarity measure 98–103, 99f, 102f trip data 114–115, 117 Outlier removement 27–28 P Parking demand 194, 194f Particulate matter emission computation 67 Pay-as-a-Package (PAAP) 153–154 Pay-as-You-Go (PAYG) 153–154 Pick-up bicycles 191 Platform as a Service (PaaS) 234 Point-based function 82 Point of interest (POI) 186 Point-to-point distance metric 80–82 Poisson distribution 145 Potential CAZs 274–276 Predictive analysis 161–162 Prescriptive analysis 161–162 Privacy issues 261 Public bicycle data 209 Public transit 18 data 22 Public transportation 11 Public transportation systems (PTS) 245 Public transport card data 208–209 Public transport network data 207 Public Transport Network Design Problem (PTNDP) 147 R Radiation type lines 167 Radio Frequency Identification Device (RFID) 190–191 Random error 215 Random forest 31 Rapid mass transit (MRT) 147–148 Recurrent neural network (RNN) 123–126, 198 Region of interest (ROI) 186 Reinforcement learning (RL) model 146 Return bicycles 191 Ride-sharing modes 181–182 Ridesharing order-dispatching problem 146 Ride-sharing potential analysis 60 computational complexity 49, 50f deep learning architecture 54 demand assessment 48 emission analysis 57–59, 59f human travel mode detection model 49, 51 matching and spatial analysis 55–57, 57f matching feasibility estimation model 51, 52–53f, 53 simulation of 49, 49f overview 48 scenario analysis 49 Ring-type lines 168 Random forest 31 Road intersection scale 253 Road network data 207 Road scale 253 Robust optimization 159–161 Rolling Time Horizon approach 144 S Sample average approximation (SAA) method 144–145, 158–159 Santiago transit system 147 Scooter sharing 18 Security technologies 236–237, 237t Segment-based function 82 Separation hyperplanes 210, 210f Sequential minimal optimization (SMO) algorithm 211 Shape-based function 82 Sharing transportation mode See Data-driven estimation, for urban travel shareability, sharing transportation mode Sharing travel 182 Shift operation 100 Significant places identification 38, 41–42 Similarity coefficient 220 Similarity function, of trajectory 82–87 Similarity measurement 220–221 Single center cities 271 Singular value decomposition (SVD) 135 Smart cards 245 data 185 Socioeconomic data 22 Software as a Service (SaaS) 234 Spatial braking 67, 69f Spatial clustering 36 Spatial entropy 195, 195f Spatial similarity 101 and dissimilarity 94–95 Spatio-temporal data preprocessing case map matching 60–72 stay location detection 35–48 travel segmentation and mode detection 48–60 map matching 33–35, 60–72 outlier removement 27–28 raw GPS data 26–27 stay location detection 29–30, 35–48 travel mode detection 31–33, 48–60 Bayesian network 32–33 hidden Markov model (HMM) 33 random forest 31 travel segmentation 30, 48–60 workflow 26–27 Standard formulation 241 Static data 150, 207 Static method 181 Station and line status monitoring problems 251 Station-based car-sharing 11 Station-level volume analysis 252 Stay location detection 29–30, 35–48 Stochastic Allocation Simulator 157 Structural similarity 102 and dissimilarity 95 Suburban lines 168 Sum and Hadamard product 133 Sum-of-Pairs distance 83 Sum of squares (SSE) 221 Support vector machine (SVM) 209–213 Sustainable urban development, MaaS case 272–276 overview 265–266 strategies 269–271, 270f macroscale 270 mesoscale 271 microscale 271 urban traffic and space 266–269 spatial structure 269 structure 267–269 T Taxi 11 Geometry 81 GPS data 185 operation data 209 sharing modes 181–182 Taxi-hailing 18 Topology-attribute matrix (T-A matrix) 92 Track and trace (T&T) data 178, 184–185 Traditional detector data 207–208 Traffic accident analysis 253 Traffic jam analysis 253 Trajectory clustering 87–90, 88–89f Trajectory data 114–116 Trajectory segmentation method 88 Trajectory similarity 79–90, 80f Transaction data 22 Transforming distance-based OD similarity measure 97–98 Transportation data types 207–209 Transportation network 186 Transportation policy 22 Transportation sensor 21 Transportation service suppliers 248–252, 250f analysis and optimization 251–252 case study 255–256, 256f object movement monitoring 249 operation status monitoring 249–251 Travel mobility similarity estimation and clustering method call detailed record (CDR) 77–78, 78f case study 103–106 CDR-based travel estimation accuracy analysis 103–105, 104–105t metro usage pattern clustering 106, 107f framework of 78f origin-destination (OD) matrix similarity 93–103, 94f image-based measure 96–97 tableau similarity measure 98–103, 99f, 102f transforming distance-based measure 97–98 volume difference focused measure 95–96 overview 77–79 trajectory similarity 79–90, 80f point-to-point distance metric 80–82 similarity function 82–87 trajectory clustering 87–90, 88–89f travel pattern similarity 90–93 travel pattern clustering 93 travel pattern expression 92 travel pattern extraction 91–92 Travel mode detection 31–33, 91 Bayesian network 32–33 hidden Markov model (HMM) 33 random forest 31 ride-sharing potential analysis 48 Travel pattern clustering 93 expression 92 extraction 91–92 similarity 90–93 Travel segmentation 30 and mode detection 48–60 Travel survey data 183 U Uber 2, characteristics big data technologies 20 integration 18 price 19 sustainability 19 overview 17 services bike-sharing 18 car-hailing 18 carpooling 17 public transit 18 scooter sharing 18 taxi-hailing 18 UbiGo characteristics big data technologies 12 integration 11 pricing 11 sustainability 12 overview 10 services bike subscribing 11 car rental 11 public transportation 11 station-based car-sharing 11 taxi 11 Under-supplied CAZs 274–276 Urban agglomerations, vs metropolitan areas 270 Urban planning analysis 253 Urban scale PM emission 72 acceleration smoothing 63–66 braking event detection 66–67 deceleration smoothing 63–66 emission analysis 67, 69f map matching 61, 63 overview 60 particulate matter emission computation 67 spatial braking 67, 69f temporal change pattern analysis 68–69 Urban traffic and space 266–269, 268f spatial structure 269 structure 247–248 V Vaite system 257 Velocity change rate (VCR) 30, 51 Visual analytics 252 Volume similarity 101, 105 W Walk distance 196, 197f Wasserstein metric 97 Whim characteristics big data technologies 14 integration 14 price 14 sustainability 14 overview 12–13 services city bike seasonal pass 13 30-day season ticket 14 pay-as-you-go 13 10-ticket 13 Whim unlimited 13 Willingness to pay (WTP) 152 Z Zimride ... for the growth of big data because they make data storage cheaper and enhance the ease of engaging with big datasets Nevertheless, the volume of big data has been at an all-time high, as is the... area both in academic and industrial aspects Though several research studies and technical reports are available, a clear link to understand big data in MaaS appears vague and fragmented In this... impacts that the MaaS has on the way we live and the changes in what MaaS stands for However, we are only at the early stages of MaaS development We would expect to see more changes appear within