Journal of Transportation Engineering, Part A: Systems

Traffic density is an essential parameter for assessing the performance of uninterrupted facilities during planning, design and operations. This study reviews and evaluates existing density estimation methods. The density estimation methods are classified into two categories: definition of concept (DOC) methods and automated methods. The traditional approach includes aerial photographic, fundamental traffic-flow equation, Edie’s, and cumulative input-output methods. The automated approach includes the occupancy method, filtering method, and integrated method. In addition, the strengths, weaknesses, opportunities, and threats (SWOT) analysis for the preceding methods is conducted. The SWOT analysis will be helpful to researchers and practitioners in selecting the appropriate density estimation method based on its relative merits. Finally, guidelines for using various density methods for heterogeneous and homogeneous traffic are presented.

Bus passenger flow is one of the decisive factors for the development of public transportation. Therefore, accurate prediction of real-time passenger flow on bus routes not only helps bus companies to make reasonable scheduling plans to meet the travel needs of passengers but also promotes the sound development of urban public transportation and reduces pollution. In this paper, we propose a secondary decomposition integration method that combines empirical modal decomposition (EMD), sample entropy (SE), and kernel extreme learning machine (KELM) to achieve a short-time prediction of bus route passenger flow. The EMD decomposes the original passenger flow data into several intrinsic mode functions, measures the complexity of the decomposed intrinsic mode functions using SE, and performs a secondary decomposition of the intrinsic mode functions with the highest complexity using EMD, followed by the prediction of the two decomposition results using KELM. The final predicted result is the sum of the two results. The model is verified by the real card-swiping data of two bus lines per minute. Each group of data has 300 data, with 80% of the data as the training set and the remaining 20% as the test set, which can predict the passenger flow per minute. The experimental results show that the short-term bus passenger flow forecasting method proposed in this paper has high accuracy and good robustness.

Abstract

Deep reinforcement learning has strong perception and decision-making capabilities that can effectively solve the problem of continuous high-dimensional state-action space and has become the mainstream method in the field of traffic light timing. However, due to model structural defects or different strategic mechanisms of models, most deep reinforcement learning models have problems such as convergence and divergence or poor exploration capabilities. Therefore, this paper proposes a multi-agent Soft Actor–Critic (SAC) for traffic light timing. Multi-agent SAC adds an entropy item to measure the randomness of the strategy in the objective function of traditional reinforcement learning and maximizes the sum of expected reward and entropy item to improve the model’s exploration ability. The system model can learn multiple optimal timing schemes, avoid repeated selection of the same optimal timing scheme and fall into a local optimum or fail to converge. Meanwhile, it abandons low reward value strategies to reduce data storage and sampling complexity, accelerate training, and improve the stability of the system. Comparative experiments show that the method based on multi-agent SAC traffic light timing can solve the existing problems of deep reinforcement learning and improve the efficiency of vehicles passing through in different traffic scenarios.

Practical Applications

This paper is devoted to research on the timing method of traffic lights at multiple intersections. The experimental results show that the method proposed in this paper can effectively improve the throughput of each intersection, the waiting time of vehicles, and the number of queued vehicles. In comparison with related algorithms, it is fully proven that the method proposed in this paper can effectively solve the problems pervasive in existing algorithms. In the actual application process, the traffic status information is obtained via interactions with real traffic environment, and the timing scheme of the traffic lights is dynamically adjusted according to the information, so as to achieve the effect of alleviating traffic congestion at the intersection.

This study aims to identify the psychological constructs that are antecedents to people’s intention to travel to a tourist destination during the COVID-19 pandemic using the extended theory of planned behavior (ETPB) and necessary condition analysis (NCA). Online survey responses collected from 1,259 participants in the United States in May–June 2021 are used for the study. We find a moderating effect of public trust towards the government on the relationship between travel concerns and intentions. Results suggest that certain levels of public trust, subjective norm, perceived benefit of travel, perceived behavioral control, and perceived knowledge of the pandemic are necessary to manifest travel intentions. We uncover that providing travel incentives and better dissemination of the pandemic-related information can potentially encourage people to regain their original travel intention that was lost due to the pandemic. Furthermore, the bottlenecks obtained using NCA show that travel intentions are more easily manifested by perceived knowledge of the pandemic versus others. This study demonstrates the application of NCA, which can be further extended to make policy-level decisions for transportation systems.

With the continuous enrichment of traffic Internet-of-Things data acquisition methods, more and more spatiotemporal data on road networks is collected in real time by various sensors and multimedia devices. The data-driven deep learning approach can make full use of real-time data from a road network to predict future traffic status. By mining the spatiotemporal relationships between road units, the ability to predict network evolutionary behaviors is improved, which provides a new method of traffic management. There are strong semantic relations between road intersections or road sections in terms of traffic evolution. Modeling the network only from a shallow spatial topological perspective ignores the important intrinsic association of the dynamic network. In this paper, we propose a semantic associative neural network (SANN) for traffic evolution analysis by modeling the propagation effects and similarity patterns between road units. Considering the inadequacy of the fixed adjacent matrix, graph convolution is used to encode the semantic features of a road network and embed them in a bidirectional recurrent neural network for sequence prediction. Finally, the experiments are conducted based on speed data sets to prove the effectiveness of the proposed method. The model achieved a well-predicted accuracy of 95.33% and 84.08% on Pems-Bay and Los Angeles data sets.

With the development of new energy technologies, dual-source trolleybuses are widely used in public transportation systems. A synergistic schedule can further improve the service quality of vehicles and optimize the passenger travel experience. This study investigated the synchronized scheduling of dual-source trolleybus networks to reduce the transfer time of passengers and the fleet size problem to reduce the operating cost of enterprises. A bi-objective mixed-integer linear programming model was developed to maximize the total synchronizations and minimize the total fleet size of dual-source trolleybus lines. A two-stage algorithm was designed to obtain multiple sets of Pareto effective solutions. Meanwhile, to demonstrate the effectiveness of the proposed method, the results of numerical examples solved by the two-stage algorithm were compared with those of a genetic algorithm (GA) and a nondominated sorting genetic algorithm (NSGA-II). A real-world case study based on the Beijing dual-source trolleybus network was studied to validate the proposed model. The results show that the model can obtain synchronized schedules. The optimization of the number of synchronizations and fleet size is obvious. The model and algorithm can be applied to a large dual-source trolleybus network.

Based on a conserved higher-order traffic flow model (CHO model), we propose a new viscous macroscopic traffic flow model taking into account the diffusion effect in traffic. The model can reasonably smooth the shock wave so that the acceleration is maintained in a reasonable range. To balance the computational efficiency and accuracy, this viscous macroscopic traffic flow model is integrated with the microscopic vehicle fuel consumption and emission models to estimate vehicle fuel consumption and exhaust emissions. The local discontinuous Galerkin (LDG) method is used to solve the viscous model, and the simulation results are inputted into the microscopic models to calculate vehicle fuel consumption and exhaust emissions. Numerical results illustrate that the proposed viscous model is reasonable and that the designed scheme is feasible and effective. Moreover, we provide concrete suggestions for controlling vehicle fuel consumption and exhaust emissions.

Electric buses are environmentally friendly with the features of zero emissions and low noise levels. Therefore, it is an inevitable trend to replace traditional fuel buses with electric buses. In the urban transportation system, the most common electric bus scheduling mode is single-vehicle-type scheduling, where crowdedness in peak hours could impair the bus service level. In contrast, a high vacancy rate in off-peak hours might lead to a waste of resources. To balance efficiency and the bus service level, this paper built an electric bus scheduling model with multiple vehicle types distinguished by vehicle configurations, such as bus capacity, battery capacity, purchase cost, charging power, and electricity consumption rate. The optimization objective was to minimize the bus system costs, including bus depreciation costs, charging costs, and congestion time costs. Then, a genetic algorithm was designed to solve the model, and the model is verified by a case study of one bus line in Nanjing, China. Compared with the single-vehicle-type scheduling schemes, including large and small vehicle types, the experimental results showed that the proposed model can reduce total bus system costs by 3.52% and 6.85%, respectively. The research results can provide references for bus companies to formulate driving plans and configure vehicles.

Turbo roundabouts, emerging as a solution to conventional multilane roundabouts’ safety and operations concerns, are prevalent in Europe and are still new to North America. However, few studies and guidelines on intersection sight distance (ISD) requirements exist. This paper focuses on the basic turbo roundabouts, examines the circulatory roadway and approach design features, and develops ISD models with North American practices. The ISD requirements were identified for the vehicle approaching and entering the roundabout, which conflicts with the circulating vehicle in the circulatory roadways, and the entering vehicle at the upstream entry approach. Analytical models were developed to identify the required lateral clearance at the central island and clear sight triangles at the roundabout quadrants to satisfy the required ISD. The analyses were completed for four typical sizes of basic turbo roundabouts per Dutch design criteria. The design guidelines are provided for the required lateral clearance at the central island and the required clear sight triangles at the roundabout quadrants. They would be beneficial in establishing landscaping restrictions and determining the roadway right of way and property impacts at the intersections during the roadway planning stage.

In the conventional human driving environment, due to poor traffic information transmission and the inability to actively control vehicles, variable speed limits (VSLs) and ramp metering (RM) strategies often fail to achieve the desired results when dealing with the merging zone of an expressway under high traffic demand. To improve traffic efficiency, an integrated control method for traffic flow in the merging area of expressways is proposed by combining connected and autonomous vehicle (CAV) active lane change (LC) technology with traditional VSL and entrance RM strategies. This method uses an upgraded cell transmission model (CTM) to estimate the state of the mainline and ramp input traffic volume and then uses the genetic algorithm (GA) to solve for the optimal combination of the mainline speed limit, number of mainline vehicles changing lanes, and ramp regulation under maximized flow in the merging zone. This integrated control strategy incorporating VSL, LC, and RM (VSL-LC-RM) enables synergistic control of the mainline and ramp traffic entering the merging zone in time-space. Finally, simulation experiments are conducted with the mainline two-lane and on-ramp single-lane expressway merging zone scenarios. The results show that compared with the VSL strategy and the strategy incorporating only VSL and RM (VSL-RM), the proposed VSL-LC-RM strategy can enhance the regulation of traffic flow, improve the flow of merging bottlenecks, and make full use of the road space-time resources.

Accurate passenger flow forecasting in urban rail transit (URT) could provide a vital reference for operators’ timely operation management. However, due to the enormous scale of the metro network, it is unwise to forecast passenger flows at the station scale individually. In this paper, a forecasting framework is proposed for network-scale forecasting tasks considering both accuracy and efficiency. There are mainly three stages in the forecasting framework. Firstly, three kinds of similarity measurements are presented regarding the adjacent similarity, geographic location similarity, and trend similarity. Secondly, three similarity graphs are formed by combining the three kinds of similarity measurements and flow time series. Thirdly, the multigraph network is applied to perform passenger flow forecasting. The experimental results indicated that the proposed method performs relatively accurately for the network-scale prediction with economic time costs. Specifically, the proposed model could acquire the feasible forecasts compared with the best disaggregate model using less than half of calculation costs, and above 10% reduction in root-mean squared error (RMSE) compared with the best aggregate model in the benchmark trials. Extensive contrast experiments were conducted to investigate the sensitivity and interpretability of the proposed model.

With the wide use of smartphone travel apps, a new hybrid transit system (HTS) that integrates fixed-route transit (FRT) and app-based demand-responsive feeder transit (DRFT) services have recently emerged in many cities over the world. App-based DRFT services are coordinated with FRT to offer fast, comfortable, and well-connected mobility services to travelers. The study investigates a new variant of HTS, termed flexible HTS (FHTS), allowing FRT to have flexible vehicle departure times and capacitated vehicle loads and DRFT to have flexible feeder stations, so as to further enhance the coordination between FRT and DRFT. The coordinated operation optimization problem of such an FHTS in a travel corridor is formulated as a mixed-integer linear programming model. A novel hybrid metaheuristic algorithm is developed to solve it. In the hybrid algorithm, a constrained compass search method is used to optimize FRT vehicle departure times, and an adaptive large neighborhood search algorithm, coupled with a tailored forward time slack algorithm, is applied to solve the DRFT routing and scheduling problems. A case study in Shucheng County, China is conducted to examine the performance of the model and solution algorithms. Sensitivity analyses on feeder station layout and vehicle size are also conducted. The results show that flexible FRT vehicle departure times and flexible DRFT feeder stations have a great potential of enhancing the coordination of FRT and DRFT services by reducing operators’ costs and passengers’ total travel time. Managerial implications are to set feeder transit stations downstream of an FRT line and to keep a reasonable station spacing.

This study evaluates different proposed merging solutions that reduce the conflict between merging vehicles and mainline traffic within a mixed traffic environment using a safety measure to see which strategy might work better than others under specific traffic conditions. The mixed traffic includes various percentages of driver-operated vehicles (DVs) and connected autonomous vehicles (CAVs). The probability of noncompliance (PNC) is selected as a surrogate safety measure to assess the strategies. A MATLAB program is developed to simulate various traffic conditions at a merging area and to calculate the PNC merging for the different merging strategies. In addition, to examine the relationship between PNC and collision frequency at the merging area, the collision data at 15 merging ramps in Ottawa were collected to examine the relationship between PNC values obtained from the simulation for the case of a full-DV vehicle fleet and no management strategy (current conditions) and actual safety performance. The results confirmed the validity of PNC as a surrogate safety measure that is correlated to expected collision frequency at merge areas. By simulating all proposed merging management strategies, the results of this study showed a general trend of decreasing PNC and, hence, improved safety performance since the CAV penetration rate increases even when no management strategy is used or under the do-nothing option. However, most merging strategies had better expected safety performance than the do-nothing option, which indicates the value of implementing a merging management strategy, especially during the period of transition from a full-DV to a full-CAV fleet.

Connected and automated vehicles (CAVs) hold great promise in enhancing transportation operations and roadway capacity. However, there is still limited guidance in the field of highway capacity analysis to help agencies account for the potential benefits. This paper investigates the impact of CAVs on the capacity of two-way-stop-controlled (TWSC) intersections and quantifies capacity adjustments under different CAV scenarios. An isolated TWSC intersection is considered in this research to analyze the CAV impacts on various intersection movement capacities under different CAV market penetration rates (MPRs) and conflict flow rates (CFRs). To model CAVs, cooperative adaptive cruise control (CACC) is considered because it directly impacts the vehicle’s longitudinal car-following behavior and therefore the capacity. Additionally, enhanced gap-acceptance behavior based on vehicle-to-vehicle (V2V) communication is included because it can benefit minor road movements via advanced notifications. This paper collects data from customized, well-calibrated TWSC simulation models constructed in commercially available software. The results show that with the increase in MPRs, CAVs can substantially enhance the capacity of the stop or yield controlled movements, and the improvement is more significant under low CFR scenarios. CAVs can also reduce the follow-up headway and thus improve the potential capacity based on the existing HCM TWSC capacity estimation method. Finally, based on the data collected from the simulation, a capacity adjustment factor (CAF) table is proposed for HCM implementation.

Rapid developments in smartphones and Global Positioning System (GPS) technology have provided a new method for travel data collection. People’s travel trajectories can be passively collected through smartphones with built-in GPS sensors and processed to infer important attributes of travel behavior such as trip destination purposes. This paper examines the feasibility of using discrete choice models and tree-based machine learning models for trip purpose inference. Using smartphone GPS trajectories and land-use data that are open-sourced for academic research, four models are developed to classify trip destination purposes into one of thirteen categories. The models include multinomial logit, mixed logit, random forest, and gradient boosting decision tree. Because the data set is significantly unbalanced with more than 50% of the trips being home or work trips, a two-stage modeling process is applied to identify home and work trips first and then subsequently to classify trips for the remaining discretionary purposes. The results show that both discrete choice models and machine learning models can achieve more than 70% prediction accuracies for home trips and work trips, but less than 30% prediction accuracies for most of the discretionary trip purposes. The discrete choice models can accurately reproduce the market share distribution of the trip purposes, but the machine learning models fail to do so. The results imply that as purely data-driven models, machine learning models may not be the best solution for trip purpose inference because they lack the theoretical background provided by microeconomics and human psychology that are essential in explaining people’s travel and activity choices.

Lane-changing recognition is an important task for advanced driver assistance systems, but is heavily challenged by poor driving habits, such as turning without using turn signals. To address this problem in this study, a lane-changing recognition method using frequency analysis was proposed. First, highest-frequency–based and frequency-bands–based methods were employed to evaluate the three behaviors of left lane changing, lane keeping, and right lane changing. To improve the recognition accuracy, the two methods were fused according to their classification advantages for different behaviors. The fused method was verified by lateral position data incorporating lane features that were manually extracted and annotated from the Next-Generation Simulation dataset. The frequency analysis framework achieved recognition accuracy of 91.8%, 97.4%, and 99.1% in 2, 1, and 0 s, respectively, before the vehicle crossed the lane line, which were significant improvements over the time-domain analysis methods. The proposed method was also validated by real-world road data with promising results.

Transportation agencies in the cold region are responsible for developing winter traffic models and verifying their sustainability to save financial and human resources while enhancing the suitability of the developed models. To do this, they operate traffic monitoring sites to collect traffic volume and loading data in their network using technologies such as permanent traffic counters (PTCs) and weigh-in-motion (WIM). None of the previous studies have conducted spatial transferability of the winter traffic models’ parameters between homogeneous and heterogeneous road segments during the winter season. This research pursues this using traffic data collected from six WIM sites in Alberta, Canada. Winter traffic models were developed for two WIM sites that serve commuter traffics. The other four WIM sites serving different travel populations besides commuter traffic were exhaustively utilized to test the developed models. The raw WIM data were aggregated into three vehicle types to develop winter traffic models by associating traffic data with climatic information. Two spatial transferability tests for the developed models were designed and carried out. The first test was conducted between the two modeling sites for which the winter traffic models were developed. The first experiment pursued a cross-spatial transferability test between homogeneous road segments. The second experiment tested the transferability of model parameters between heterogeneous road segments that represent a different road function other than commuter type. The models’ parameters developed for the two commuter segments were transferred to the other four sites to test their spatial transferability. This research has demonstrated that the winter traffic models developed for the roads serving one specific travel population can be transferred with high accuracy to homogeneous and heterogeneous road segments. It revealed that a more suitable model structure could be selected for each site and vehicle class, considering the accuracy of the test results. This research contributes to planning and designing traffic monitoring or weighing site deployment to save financial and human resources.

To alleviate traffic congestion, an improved car-following model for connected vehicles is proposed in this paper by considering backward-looking effect and motion information of multiple vehicles. The backward-looking effect is common in traffic and motion information of multiple vehicles and is beneficial for alleviating traffic congestion. The linear and nonlinear stability of the proposed model is analyzed, and the stability condition and modified Korteweg-de Vries equations of the proposed model are derived. The theoretical analysis results prove the effectiveness of the proposed model in alleviating traffic congestion and improving the stability of the traffic system. Further, numerical simulation is designed to verify the promoting effect of the introduced parameters on the traffic stability and to test the effect of the improved model on large-scale car-following queues. Finally, the next-generation simulation (NGSIM) data sets are used to calibrate the parameters of the improved model. The results show that the improved model can effectively avoid traffic congestion and enhance the stability of traffic flow. The improved model can be used as active safety technology to prevent collision accidents or as a car-following strategy in driverless algorithms.

Based on the traffic conflict techniques, a microscopic safety analysis framework is presented to study traffic conflicts in the freeway exit ramp diverging area (FERDA) during the maintenance period. A manual–software combination data processing approach was used to extract trajectory data from aerial videos and identify actual conflicts. Simulation conflicts were obtained from the VISSIM simulation model. Firstly, the conflict features of maintenance and normal operation periods were compared. Subsequently, conflicts were classified as lane-change conflicts or rear-end conflicts to investigate their spatiotemporal characteristics during the maintenance period. Finally, a random parameters multinomial logit model was developed to study the influencing factors and heterogeneity of conflicts. The results revealed that the FERDA has a higher collision risk during the maintenance period compared with the normal operation period. Additionally, lane-change conflicts have a higher occurrence probability than rear-end conflicts, but a shorter duration. Besides, the conflict density heat map showed that the lane-change conflicts and rear-end conflicts are densely distributed at 0–150 m and the middle lane of the FERDA, respectively. The model results indicated that the factors that significantly influence conflicts include velocity, acceleration, type, position, and destination of vehicle, the heavy vehicle percentage, total number of vehicles in the FERDA, and the velocity and destination difference between vehicles. Among these factors, the type of vehicle, destination difference, and the position of vehicles (at 150–270 m of the FERDA) were found that have significantly heterogeneous impacts on the two types of conflicts. This study is expected to provide safety analysts with an effective analysis method for evaluating the risk of the FERDA during the maintenance period. Additionally, the concluded influencing factors can potentially help develop different and effective safety countermeasures for different types of conflicts.

Urban public security incidents are prone to occur. Better understanding of pedestrian abnormal behavior and trajectory in crowded places is conducive to crowd management and safety monitoring. A novel pedestrian abnormal behavior detection model (PABDM) is proposed to identify crowd behavior under abnormal scenarios. This model originated from a multiscale fusion you only look once (YOLO) version 3 (V3) algorithm and was trained using the PASCAL visual object classes (VOC) in combination with an abnormal pedestrian data set (APD), denoted as VOC+APD. Compared with YOLOV3-VOC, single-stage detectors (SSD)-VOC, and $\mathrm{SSD}\text{-}\mathrm{VOC}+\mathrm{APD}$, the proposed model has notable advantages in prediction accuracy and detection efficiency. The results show that the network loss function of the model tends to be stable after 500 epochs, and its detection accuracy is 6% higher than the average accuracy of the compared models. This proposed model also effectively solves the problem of missing detection caused by edge target, fuzzy target, and small target in abnormal state human detection. The research results are of great significance for real-time crowd monitoring in complex scenes.

Several studies have measured the minimum operating speed on horizontal curves to model the operating speed to assess the geometric design consistency. Most of these studies approximated equal lengths of deceleration and acceleration in the operating speed profiles for the curves and assumed the minimum operating speed position at the midpoint of the curve. In contrast, a few recent studies showed different percentages of deceleration lengths on the curve and measured the minimum operating speed at the deceleration end on the curve to model the operating speed. A defined pattern of the operating speed profile on the horizontal curve was not reported in the previous studies and therefore presents opportunities to determine the patterns of the operating speed profiles on curves. In this study, the operating speed profiles of different drivers for the given features of the horizontal curve were studied, and the clustering technique was used to categorize the different patterns in the operating speed profiles on horizontal curves. The optimal number of clusters was determined using four methods: silhouette, elbow, gap statistic, and NbClust function. The different patterns observed from the clustering results are as follows: (1) complete deceleration on the curve, (2) complete acceleration on the curve, (3) deceleration length slightly greater or lower than acceleration length, and (4) longer deceleration/acceleration lengths followed by shorter acceleration/deceleration lengths, respectively. The study results imply that all operating speed profiles are not symmetric around the midpoint of the curve (MC), and the group of drivers exhibited defined patterns of the operating speed profiles on the curves. This study helps in understanding the different patterns of operating speed profiles exhibited by the drivers and the measurement of the minimum operating speed at the deceleration end to model the operating speed to assess the geometric design consistency.

Road crashes cause significant traffic delay, which can bring unnecessary financial losses. The objective of this study is to predict the level of delay caused by crashes (LDC) and discuss significant risk factors. To ensure the efficiency and accuracy of prediction, an improved stacking model was developed using Texas crash data of 2020. The first layer integrates seven base classifiers and the second layer tests three classifiers with different advantages. To improve and simplify the stacking model, three state-of-the-art methods—Bayesian hyperparameter optimization (BO), multiobjective feature selection (FS), and ensemble selection (ES)—were used. First, the hyperparameters and the least and most effective features were selected for each base classifier by BO and FS, respectively. Then ES, considering diversity and performance, selects the least base classifiers to reduce the input of the second layer. Finally, permutation feature importance was used to interpret the best stacking model. The results indicate that the stacking model achieves superior performance on four indicators: recall, $G$ mean, F1 score, and area under the receiver operating characteristic (ROC) curve (AUC-ROC). FS significantly improves the efficiency of the stacking model and ES obtains a simplified stacking model without significantly reducing performance. In addition, the combination of the two methods (FS and ES) tends to achieve the best performance, and six risk factors have the greatest contributions in prediction using permutation feature importance. The prediction of LDC and the analysis of the main contributing factors help road managers respond to the rescue strategies to mitigate traffic congestion caused by crashes in a timely manner, thus minimizing economic losses.

Abstract

COVID-19 had serious repercussions on public transportation throughout the USA. The aftermath of the peak of the crisis marked the path towards a slow and gradual recovery characterizing the shift to a new normal. Given the limited information on the recovery trends of public transportation, this paper compares the actual ridership and bus supply data for the years of 2019 and 2020 to study the timeline impacts of the pandemic on the bus system of the mid-sized city of Syracuse, NY. A data-driven analysis is presented across the city’s bus routes, university bus routes, and categorical bus stops. Various census tract socio-demographic data are also correlated with passenger activity changes and mapped using ArcGIS. The findings show that overall bus ridership in 2020 fell by 70%, on average, during the three months that followed the onset of the pandemic. Since the lifting of the initial restrictions, concerns about using public transportation had partially been alleviated; however, passengers remained reluctant with ridership decline stabilizing at approximately 55% during the last four months of the year. While bus lines serving the university area, which houses a high percentage of youth, were severely affected by the pandemic, passenger activity near hospital stops were less affected and those near major supermarkets/hypermarkets seemed unaffected, showing a surge especially in the two months that followed the onset of the pandemic. Passenger activity at census tracts having low poverty levels mostly located on the outskirts of the city of Syracuse were the least affected tracts in the last six months of 2020. It is anticipated that the insights presented will help service planners in preparing for similar future events by better understanding what stops and routes are deemed essential during a public health crisis and how the socio-demographics impacted the recovery after restrictions were removed.

Practical Applications

The early stages of the pandemic prompted transit operators and authorities to impose restrictions on public transportation to limit transmission of the virus and ensure the safe commute of essential workers. This paper studies actual ridership and bus service data in connection with sociodemographic data to understand the impacts of the pandemic on the sole public transportation system in Syracuse (NY). The study focuses on four time periods: pre-pandemic, onset of the pandemic, transition period, and recovery period. While bus lines serving the university area were severely affected by the pandemic, passenger activity near hospital stops were less affected, and those near major supermarkets/hypermarkets seemed unaffected, showing a surge especially in the two months that followed the onset of the pandemic. Ridership in areas with low poverty levels that are mostly located on the outskirts stayed at high levels in the months that followed the outbreak. The insights presented in this paper can help public transportation planners better prepare for similar disruptive events in the future as they can take targeted steps by studying what stops, routes, and sociodemographic regions were deemed essential in this case study and how the network might recover after restrictions were removed.