Keywords

1 Introduction

Intelligent Public Transportation Systems (IPTS) have been developed for decades in urban areas worldwide. The purpose of building IPTS is two-fold: (1) to assist transportation staffs to manage service and maintain the performance of transportation networks; (2) to help users plan and execute trips with ease and satisfaction [1, 2]. Paying special attention to the user side of IPTS, this paper examined Shanghai travelers’ behavior journey while taking public transportations. We adopted the approach of scenario-based modeling [3] and investigated users’ behaviors and perceptions in each phase of the trip, i.e., before/during/after the trip. We aimed to build a comprehensive user journey map for Shanghai’s IPTS, and to distill some key factors which affect users’ behavior patterns and perceptions. Findings of the study can assist IPTS developers and designers to improve Shanghai IPTS’s facilities so that better user experience can be attained.

As a metropolitan, Shanghai has relatively advanced IPTS. Real-time schedules are accessible at all metro stations, and also available at increasing number of bus stations, as well as a few mobile apps. Mobile payment is widely supported, travelers are able to use smart phones to either purchase tickets or add balance at vendor machines. If people choose to go to the ticket-less way, they can even scan phones on ticketing machines directly, without bothering taking out physical tickets or transportation cards. Complementing bus and metro, vehicle sharing (bike and e-car) and ride sharing (Uber equivalence) are easily assessable, helping travelers to arrive the stations and to complete the “final mile”. Moreover, trip planning tools, which support not only unimodal travels, but also multimodal ones, are assisting travelers to familiarize with their trips in advance. Some software can also assist travelers during the trip, such as showing which station they are at, and remind passengers not to miss transfer point or destination.

To better understand people’s usage of IPTS in the city of Shanghai, we paid attention to not only the primary public transportation modals, i.e. bus and metro, but also any complementary modals which fulfill the trips, such as vehicle sharing and ride sharing services. In addition, we examined trips with familiar routes and unfamiliar routes as distinct scenarios. For unfamiliar routes, travelers require to allocate more planning time in advance. Though trip planning is not a novice topic, most studies approach this topic from technical aspect [4, 5], while left the user experience side not getting sufficient attention. For familiar routes, the planning phase is less intense, yet travelers still need to manage their trips properly so as to avoid unexpected frustrations. For both route types, IPTS can provide helpful services and tools, but the question is how people are making use of these services and to what extent are these intelligent services really solving problems. To answer such questions, the current study adopted combined research methods, including field observation, diary entries and in-depth interviews, to investigate users journeys while using IPTS in Shanghai.

2 Literature Review

Intelligent public transportation has been studied for decades. Existing literature on this topic can be generally divided into three categories: (1) infrastructure oriented, (2) integrating user aspect, but still more technology related, and (3) user-centric. The first category deals with the foundation of intelligent public transportation system. It reviews the architecture and technologies which enable IPTS [1, 6]. The possibility of applying new technologies to enhance the systems are also discussed in these literatures, for example using Zigbee and NFC for better data transit [7, 8], and adopting radio-frequency identification technology or mobile communication network to track passenger flow [9, 10]. These studies mostly discuss IPTS from architecture-perspective, yet the user side is not covered.

The second category of relevant literature embraces the interaction between users and technologies, and covers various aspects throughout the journey. For example, during the trip planning phase, Földes and Csiszár proposed a personalized route planning model which takes users’ personal preferences into account [11]. Optimizing solutions for multimodal journey planning are also discussed in prior studies [12, 13]. Besides, time estimation is a topic that has been studied intensively. Simeunović et al. worked on an algorithm that aim to yield better transfer time estimation [14]. Pagani et al. put forward a system which not only shows estimated arrival time, but also the cumulated probability for arriving on time [15]. Moreover, some other studies pay attention to the constant changing environment while travelling. Systems have been designed to recognize and reconstruct travelers’ surrounding circumstances, and offer them with precise contextual instructions and suggestions [16, 17].

Nonetheless, though the second category of studies make contribution to the user experience of IPTS by investigating optimized algorithms and models, these studies mostly take the technical perspective, paying more attention to technical difficulties and corresponding solutions. Users’ behaviors and perceptions about IPTS are barely mentioned in these studies, but are the main topic of the third category of relevant literature. Digmayer et al. analyzed travelers’ activities by using a scenario-based approach [3]. They constructed personas and scenarios for unimodal and multimodal travel chains, and identified travelers’ information needs at each phase of the journey. By identifying these user needs, travel information apps can be better designed with proper content and functions. Islam et al. examined how travelers are using, and what factors drive their usage of ubiquitous real-time passenger information tools [5]. They claim that better understanding user behavior of using such tools can help to clarify technological areas that are necessary for more investments, and assist to make IPTS more effective.

The current study will follow the direction of the third category of relevant literature, i.e. the user-centric approach. The user side of IPTS has not been studied thoroughly. More work need to be done in this area to assist people in relevant fields to better understand travelers’ behaviors and perceptions around IPTS, which can be helpful to technical-oriented studies as well.

3 Methodology

We employed a combination of methods for the current study. Field observations, diary recordings and in-depth interviews were conducted consecutively, leading the researchers to gain a better understanding about user’s behavior patterns and perceptions around IPTS related tools and services throughout their journeys.

3.1 Field Observation

Field observation was carried out as the first phase. This phase purpose to assist the researchers to familiarize with the current implementation of intelligent services in Shanghai’s public transportation system, so as to construct a more accurate and up-to-date context for later examinations. Also, by doing field observation, the researchers were able to gain a general sense about how Shanghai citizens are using and thinking about their intelligent public transportation system.

We conducted observations at 5 bus stations and 5 metro stations, spread across 2 major districts in the city of Shanghai, i.e. YangPu District and HuangPu District. At each station, we spent 30 min, paying special attention to these stations’ implementation of intelligent transportation services, including real-time schedule boards, mobile payment portals, availability of nearby sharing bikes and e-cars, as well as ride-sharing pick-up areas. We also focused on people’s behaviors and perceptions around these services, taking notes about people’s usage of intelligent tools and conducting random short interviews. Our field observation covered both rush and non-rush hours. All observed data were recorded properly and prepared for further analysis.

3.2 Diary Recording

Following field observation, 10 participants were recruited with snowball sampling technique to write diary entries. All participants take public transportation for daily commute in Shanghai, and have some experience of using intelligent transportation services. Participants aged from 20–40, as suggested by a prior study that people at this age group are the most likely to use intelligent public transportation tools [1]. Also, because significant gender difference about using intelligent public transportation service was not expected [1], the gender ratio of the current study was set to 1:1.

Participants were required to write diary entries for 4 days, including 2 workdays and a weekend. They needed to cover trips on both familiar and unfamiliar routes during these 4 days. Content of diary contain basic trip information, including date, purpose of trip, starting point, destination, departure and arrival time, as well as means of transport, see Fig. 1. Besides, usages of intelligent public transportation services were also recorded, see Fig. 2. Participants took notes about their behaviors and perception about using trip planner, travel assistant (assisting during the trip), real-time schedule board, mobile payment, biking-sharing, e-car sharing and ride-sharing, etc. Completed diary entries were sent to researchers on daily basis via the Internet.

Fig. 1.
figure 1

Diary sample – trip information section

Full size image
Fig. 2.
figure 2

Diary sample – usage of IPT service section

Full size image
Fig. 3.
figure 3

User journey map for in-city public transport with intelligent services

Full size image

3.3 In-Depth Interview

All the participants who wrote diaries were also invited to take part in in-depth interviews. The interviews were conducted either face-to-face or via video chat software, and typically lasted for 40–60 min. Though interview script was prepared in advance, when carrying out these sessions, we did them in semi-structured format and modified questions based on each individual’s diary entries. Participants were encouraged to elaborate on their behaviors and feeling at different stages throughout the whole trip, to explain how and why they choose to use certain intelligent transportation services, and to propose where they think relevant tools can be added to improve their experience of the journeys. All the interviews were voice recorded and transcribed for further analysis.

3.4 Data Analysis

During the phase of data analysis, we adopted a scenario-based approach [3], which allows researchers to closely examine the activities of users through each step of their public transportation journeys. We divided the journeys into the following steps: before leaving, on the way to transport station, at the station, on vehicle, transfer, at the final station, leaving the final station. For each step, we analyzed travelers’ behaviors, perceptions and needs, so as to reconstruct the travel chains, and to build a comprehensive scenario-based user journey map. Though the scenario and persona which represented in the journey map are hypothetical, they actually originated from coded qualitative data from our field observation, diary entries and in-depth interviews.

4 Result

4.1 Persona

To construct scenario-based journey maps, persona first need to be created. By creating persona, it would be easier for researchers to build empathy and think from the perspective of travelers. Also, persona help the researchers to delve deeper into the meanings behind users’ activities, and probe into travelers’ needs throughout each stage of the trip. According to the profile of participants who took part in our study, we built a hypothetical female persona with age of 28, whose name is Lizhen Wang. Lizhen is a game operating manager based in Shanghai. She takes public transportation for daily commute and other in-city travels, and is a heavy smartphone user. Besides, Lizhen has very poor sense of direction.

4.2 Scenario 1. Taking Metro for Unfamiliar Trip

Lizhen was invited to a friend’s apartment where she had never been. Her friend’s place was at another district in the city of Shanghai. It would take around 1.5 h to go there with public transportation services. Lizhen took a metro-based route, and transferred metro lines once during the trip. A few intelligent transportation tools were assisting Lizhen to complete the trip.

Before Leaving.

When planning the trip, Lizhen used a mobile trip planner. The planning activity was executed twice: first happened right after the invitation because she wanted to have a general sense about how to go there; Lizhen used trip planner again on the day before departure, which purposed to confirm the route. Among all suggested routes, Lizhen preferred metro-based ones, which in her mind were more reliable. Based on prior experience, Lizhen was quite confident about the estimated travel time for metro-based routes, yet she would still prepare extra 20 min in case unexpected situation happened. Lizhen hoped the trip planner could be more personalized and suggest routes according to her own preference. Lizhen barely choose to transfer between metro and bus, yet such routes were suggested to her quite often.

Way to the Initial Station.

Lizhen did not want to spent too much time on the way to the initial station, so she rode a sharing bike for the “first mile”. Since there were not plenty sharing bikes around Lizhen’s apartment, she spent some time looking for one. When arrived metro station, she parked the bike at a specific area which is very close to the station. No navigation was needed at this stage.

At the Initial Station.

Lizhen used mobile payment to enter the station. Before choosing a metro direction, she looked at her trip planner app for a double check. The metro platform has real-time schedule e-boards, but Lizhen did not look at it because she knew the metro would come every 3–5 min. She would look at the e-board when she had waited longer than the expected.

On Vehicle.

In order to get information about where the train was at, Lizhen depended on the train’s announcement as well as the electronic map. She did not use mobile travel assist for this information for the reason that her phone was occupied for entertaining activities, also because she concerned that mobile geo-location detection became inaccurate underground.

Transfer.

Lizhen followed signs in the metro station to find the place where she would take the next train. To make sure taking the correct train, Lizhen looked at her trip planner again before choosing a train direction. Again, she did not look at the real-time schedule e-board, because she knew the next train would come soon.

At the Final Station.

The final metro station was still relatively far away from Lizhen’s destination, so Lizhen booked ride-sharing service when she arrived the final station. She met the driver at a specific station exit as they communicated via phone.

Leaving the Final Station.

Lizhen shared her location with her friend when she was in the ride-sharing car, so that her friends could know where she was and when she would arrive.

4.3 Scenario 2. Taking Bus for Unfamiliar Trip

Lizhen needed to visit a client. The client’s office was at an area that Lizhen had never been to, and that place could not be easily accessed via metro. Lizhen took bus for this trip. One transfer need to be made during the trip. Navigation tool played an important role assisting Lizhen to arrive on time.

Before Leaving.

Lizhen always prefers metro because of its reliability. However, for this trip, taking metro would consume twice as much time as taking buses. With respect to time, Lizhen chose bus as the means of transport. Also, since Lizhen was not confident about the trip planner’s estimated travel time for bus travels, she prepared extra 40 min.

Way to the Initial Station.

Lizhen had never took that bus line before, so she used mobile navigation to help her look for the bus station. Though the navigator led Lizhen to the correct place, Lizhen did not recognize the bus station at the first place as that was a small station. Lizhen was hoping the navigator could show a picture of the bus station which might help people to recognize it.

At the Initial Station.

Once arrived the station, Lizhen opened the bus schedule app to find out the estimated bus arrival time. Based on prior experience, the estimated arrival time is not always accurate, but it could gave Lizhen a general sense about how long she had to wait. Besides estimated time, this app also indicated how many stops remaining for the bus to come. This is a feature that Lizhen liked particularly because it added some certainty to bus travels.

On Vehicle.

Because the announcement on bus could be inaccurate and misleading, when on the bus, Lizhen used navigator to track her own location. The navigator showed which station she was at, and how many stations remaining before getting off. Nonetheless, since phone was used for navigating, Lizhen could not do other entertaining activities such as watch streaming videos. Lizhen hoped the navigator could sent her a push notification few stops before getting off, so that she did not have to open the navigator for the whole time.

Transfer.

Lizhen got off the first bus at station A, while she needed to take the second bus at station B. Once getting off the first bus, Lizhen used navigator to look for station B, and also opened bus schedule app to check the second bus’s arrival time. Though this time Lizhen did not spend too much time on transfer, she actually could not know how long it would beforehand, neither did she trust the estimated transfer time from trip planner.

At the Final Station.

Lizhen looked at the bus stop name to confirm that she arrived the correct station. Then she found a sharing-bike parking site near the station, so she took one and headed for destination.

Leaving the Final Station.

Lizhen needed navigation for the “final mile”, but she was riding a bike without a phone mount. Therefore, Lizhen had to stop occasionally to look at her phone to make sure that she was on the right way.

4.4 Scenario 3. Familiar Trip

For daily commute, Lizhen can either take bus or metro. When taking bus, there is no need for transfer, and usually takes less time. When taking metro, Lizhen needs to transfer once.

Before Leaving.

Lizhen checks estimated bus arrival time via mobile app, and leave home when the bus is coming soon. No intelligent transportation tools are used if she choose to take metro to workplace.

Way to the Initial Station.

Lizhen walks to bus/metro station from home. If she finds a sharing bike on the way, she will use it. However, she will not look for sharing bike purposefully, because looking for a well-functioning bike may consume more time than walking to the station.

At the Initial Station.

If taking bus, Lizhen looks at estimated bus arrival time from phone once she arrives station. If taking metro, she uses transportation card to enter the station. Lizhen does not use mobile pay during rush hours because mobile pay is not reliable sometimes, which can cause frustration and delay for commute.

On Vehicle.

Lizhen is very familiar with the daily commute route, so no travel assistants are needed during this stage.

Transfer.

When taking metro for commute, Lizhen needs to transfer once. She will look at the estimated train arrival time on the platform to make sure that she can arrive office on time. The estimated arrival time for metro are usually accurate.

At the Final Station.

Lizhen will try to find a sharing bike to reduce the time spent on the “final mile”. Nonetheless, there are not always enough sharing bikes parked around the final bus/metro station, so Lizhen usually left home a bit earlier in case she has to walk the “final mile”.

Leaving the Final Station.

Lizhen is familiar with the route from the final station to office, so no mobile navigation is needed.

5 Discussion

5.1 Travelers Heavily Rely on Intelligent Transportation Tools

With the existence of trip planner and navigator, travelers do not have to memorize unfamiliar routes and are able to refer to relevant mobile apps for confirmation at any time. As one participant pointed out in the interview that “I like to keep it (a mobile app merging the function of both trip planning and navigating) open for all the time. Every time I need to make a decision (such as choosing a side at the metro station, or deciding whether to get off a bus), I’ll look at it. I just cannot remember those information, neither do I want to memorize it.” Besides trip planning and navigating software, other intelligent transportation tools are also frequently used throughout various stages of the trip, for instance, scanning smartphone for mobile payment when entering/exiting metro stations or getting on/off buses, and unlocking sharing bikes for the trip’s first/last mile. For participants in the current study, their usage of intelligent public transportation tools covers the timespan of the whole trip, and they have built high dependency on these tools.

These tools, on one hand, are assisting travelers to complete their trips with more ease, but on the other hand, are also creating an anxiety about ensuring their smart phones are functioning well throughout the whole trip. Battery is the prerequisite for a phone’s well-functioning, and is also a factor that travelers concern a lot: “I barely used my phone on metro though I really wanted to, because I still need to use the phone to exit the metro station, and also use it to unlock a sharing bike. I was just hoping my phone do not die on the way.” In addition to low phone battery, poor phone signal is another factor that can bother travelers. When travelling to an area where phone signals were barely receivable, it could be desperate. Travelers heavily rely on travel assisting tools, while these tools only work when the smartphones are functioning well. If foreseeing phones’ incapability, participants in our study would choose alternate method in advance, such as using physical transportation cards for payment, or bring supplementing devices, for example portable power banks.

5.2 Travelers Want More Certainty During the Journey

Travelers tend to avoid transferring between metro and bus, or bus and bus, because of much uncertainty. As a participant mentioned in the interview “oftentimes I need to go to another bus station for the next part of my journey. This worries me because first I am not confident about my ability to find that bus station without too much effort, and second I do not know how long I need to wait for that bus.” Neither are travel assisting tools dealing well with such uncertainty caused by transferring with buses. Though researchers are working on algorithms to estimate transfer time happened in public transportation [14], according to our participants, the performances of existing trip planners, with regard to estimating time for transfer with buses, are not satisfying. Participants never know how long it will take until they are actually doing the transfer. For such reasons, travelers are more likely to avoid paths with bus-transfer even tripper planners are indicating that these paths are taking the least time.

Another uncertain situation that travelers tend to avoid is using unreliable mobile payment methods during rush hours. Mobile payment methods are getting iterated and optimized gradually, however, at the current stage, the hardware and software that some traveler use do not perform well consistently. Problems that travelers encounter include slow loading time, QR code not being identified, no low balance alert, malfunction station gate, and etc. These problems can happen to travelers occasionally and make them late for commute or blocking other people who are waiting to enter the stations. For such reason, a more reliable alternate, such as physical transportation card, are still used by mobile payment users during rush hours. Nonetheless, participants also mentioned that unreliable issues are more likely to happen to QR code-based mobile payment methods, instead of NFC-based ones. Therefore, it can be expected that if NFC-based mobile payment become wider adopted or QR code-based methods resolve the reliability issue, mobile payment may play an important role during rush hours.

5.3 Travelers Trace for Smoothness While Travelling

The reason of travelers’ avoiding uncertain situations is partly due to their pursuit of smoothness during the trip. Both unpredictable bus-transfer and unreliable QR code-based mobile payment can make travelers feel trapped in certain stage of the journey. On the contrary, transfer within metro stations and pay fare with reliable methods are less likely to yield unexpected troubles and afford smoother travel experience, therefore receive popularity. In addition to transfer and payment, another issue that closely relates to travel smoothness is the designated parking area of sharing bikes. Though sharing bikes are making contribution to resolving the problem of “first/last mile”, assisting travelers to arrive destinations more efficiently, relevant experience is not always as satisfying as expected. As our participants mentioned, “sometimes I really have to spend a lot time looking for a bike, it makes me upset” and “a few times, I have to park the bike kind of far away from the station, and then walk there, it is like extra walking distance for me”. It is reasonable to expect that the travel of “first/last mile” can be smoother if more sharing bike parking areas are designated around transportation stations, residential areas, as well as business districts. Nonetheless, such parking lot designation also need to ensure not causing other problems, such as blocking the pedestrian lane. Various stakeholders need to be involved to work out a solution that suits best for all kinds of travelers, as well as the city overall.

5.4 Travelers Are Looking for More Human-Oriented Experience

Though existing trip planners and navigators allow users to modify paths by choosing fewer time/transfer/walking, or bus/subway first, participants in the current study are hoping the intelligent tools can consider even more experience-related factors, such as weather and condition of roads. For people taking public transportation, weather is a crucial factor to be taken into account, as different weather can dramatically affect travelers’ choice of transport, the travel time, as well as the overall travel experience. However, to our best knowledge, current travel assisting tools have not taken weather into consideration, and make traveler feel that “it only plans routes based on some mysterious numbers while I would like it to consider more about my comfort.” Some trip planners will suggest travelers to wait at an unsheltered bus station during a heavy raining day, which according to a participant “is very unhuman”. Besides weather, road condition is another “human factor” that travelers care about. As a participant pointed out “I will not always follow its walking navigation, I’ll look at the general direction and then choose my own path. The navigator sometimes points me to some unclean and narrow streets, which may take less time, but I definitely do not want to go that way”. For travelers, they hope the travel assisting tools can regard them more as “human instead of data points”, and give travelers’ comfort and safety higher priority while planning routes for them.

6 Conclusion

Intelligent public transportation systems are providing helpful services for travelers to complete their journeys with more efficiency and better experience. It seems that some travelers have built very high dependency on intelligent travel assisting tools throughout various phases of their journeys. However, such high usage of relevant tools also causes anxiety about the well-functioning of smartphone. Besides, travelers are looking for journeys with more certainty and smoothness, while existing IPTS still have some incapability in addressing these objectives. Travelers also hope travel assisting tools can take more human factor, such as the degree comfort and safety, into consideration, so that people can regard them are more thoughtful and caring assistants, instead of tools without much warmth.