Pub Date : 2025-12-31DOI: 10.1016/j.jcmr.2025.100105
Daniel Kerger, Heiner Stuckenschmidt
Public mobility sharing systems are an important component of sustainable transport, particularly for last-mile travel. However, analysing trip patterns using open standards such as GBFS can be challenging due to vehicles frequently being assigned new identifiers and missing GPS trajectories, preventing a detailed tracking. To overcome this limitation, we present a machine learning pipeline that retrospectively predicts trip destinations within this circumstances—making it possible to partially recover travel patterns for GBFS data.
Our approach involves a three-step prediction pipeline: (1) candidate generation and reduction using spatial–temporal filtering; (2) multi-target regression via XGBoost to estimate destination coordinates; and (3) selection of the best-matching candidate. Our approach achieves an average accuracy of 77% across five German and 74% across five international cities within a tolerance of 500 metres. Compared to existing approaches, our method improves prediction accuracy by an average of 20% over methods that also do not use user-specific or GPS trajectory features.
These results demonstrate the feasibility of accurately predicting destinations in shared mobility despite rotating vehicle identifiers and missing trajectory data, thereby supporting improved system analysis and planning.
{"title":"Multi-operator free-floating GBFS trip destination prediction in public mobility sharing systems","authors":"Daniel Kerger, Heiner Stuckenschmidt","doi":"10.1016/j.jcmr.2025.100105","DOIUrl":"10.1016/j.jcmr.2025.100105","url":null,"abstract":"<div><div>Public mobility sharing systems are an important component of sustainable transport, particularly for last-mile travel. However, analysing trip patterns using open standards such as GBFS can be challenging due to vehicles frequently being assigned new identifiers and missing GPS trajectories, preventing a detailed tracking. To overcome this limitation, we present a machine learning pipeline that retrospectively predicts trip destinations within this circumstances—making it possible to partially recover travel patterns for GBFS data.</div><div>Our approach involves a three-step prediction pipeline: (1) candidate generation and reduction using spatial–temporal filtering; (2) multi-target regression via XGBoost to estimate destination coordinates; and (3) selection of the best-matching candidate. Our approach achieves an average accuracy of 77% across five German and 74% across five international cities within a tolerance of 500 metres. Compared to existing approaches, our method improves prediction accuracy by an average of 20% over methods that also do not use user-specific or GPS trajectory features.</div><div>These results demonstrate the feasibility of accurately predicting destinations in shared mobility despite rotating vehicle identifiers and missing trajectory data, thereby supporting improved system analysis and planning.</div></div>","PeriodicalId":100771,"journal":{"name":"Journal of Cycling and Micromobility Research","volume":"7 ","pages":"Article 100105"},"PeriodicalIF":0.0,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926811","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-29DOI: 10.1016/j.jcmr.2025.100100
Madeleine Bonsma-Fisher , Moreno Zanotto , Kathryn L. Walker , Samuel Benoit , Sara F.L. Kirk , Meghan Winters
Pedal Poll/Sondo Vélo, Canada’s national volunteer bicycle count, completed its fourth count year in 2024. We analyzed this large, crowdsourced dataset of 204,584 people on bikes counted across 64 Canadian communities over four years to examine trends in cycling rates (people on bikes/hour) and their associations with cycling infrastructure type and accessibility to destinations. We classified the infrastructure at count sites according to the Canadian Bikeway Comfort and Safety (Can-BICS) classification system using Google Street View imagery and linked count sites to accessibility to destinations using national Spatial Access Measures (SAM) data. We used generalized linear mixed models to model the relationship between bicycle counts and the count year, count time of day, infrastructure at count sites, and accessibility to destinations, and included random effects for repeated sampling at the same count sites over time. We found that, relative to sites with no cycling infrastructure, medium and high comfort cycling infrastructure was associated with 55% and 105% higher cycling volumes respectively. Similarly, a 1 interquartile (IQR) increase in accessibility to destinations was associated with 65% percent increase in cycling volume. Relative to weekday morning counts, weekdays from 4–6 pm were associated with 53% higher cycling volumes, and weekends from 12–2 pm were associated with 28% higher cycling volumes. We did not see a change in the rate of people cycling over time at sites with medium or high comfort cycling infrastructure, but for the 112 count sessions at sites with low comfort infrastructure, each successive year was associated with a 12% decrease in cycling volume. These findings show that safe, high-comfort cycling infrastructure and accessibility to destinations are both associated with higher rates of cycling, and they highlight the value of a growing volunteer-collected dataset for advancing evidence on cycling in Canada.
{"title":"Cycling rate trends from Canada’s national volunteer cycling count","authors":"Madeleine Bonsma-Fisher , Moreno Zanotto , Kathryn L. Walker , Samuel Benoit , Sara F.L. Kirk , Meghan Winters","doi":"10.1016/j.jcmr.2025.100100","DOIUrl":"10.1016/j.jcmr.2025.100100","url":null,"abstract":"<div><div>Pedal Poll/Sondo Vélo, Canada’s national volunteer bicycle count, completed its fourth count year in 2024. We analyzed this large, crowdsourced dataset of 204,584 people on bikes counted across 64 Canadian communities over four years to examine trends in cycling rates (people on bikes/hour) and their associations with cycling infrastructure type and accessibility to destinations. We classified the infrastructure at count sites according to the Canadian Bikeway Comfort and Safety (Can-BICS) classification system using Google Street View imagery and linked count sites to accessibility to destinations using national Spatial Access Measures (SAM) data. We used generalized linear mixed models to model the relationship between bicycle counts and the count year, count time of day, infrastructure at count sites, and accessibility to destinations, and included random effects for repeated sampling at the same count sites over time. We found that, relative to sites with no cycling infrastructure, medium and high comfort cycling infrastructure was associated with 55% and 105% higher cycling volumes respectively. Similarly, a 1 interquartile (IQR) increase in accessibility to destinations was associated with 65% percent increase in cycling volume. Relative to weekday morning counts, weekdays from 4–6 pm were associated with 53% higher cycling volumes, and weekends from 12–2 pm were associated with 28% higher cycling volumes. We did not see a change in the rate of people cycling over time at sites with medium or high comfort cycling infrastructure, but for the 112 count sessions at sites with low comfort infrastructure, each successive year was associated with a 12% decrease in cycling volume. These findings show that safe, high-comfort cycling infrastructure and accessibility to destinations are both associated with higher rates of cycling, and they highlight the value of a growing volunteer-collected dataset for advancing evidence on cycling in Canada.</div></div>","PeriodicalId":100771,"journal":{"name":"Journal of Cycling and Micromobility Research","volume":"7 ","pages":"Article 100100"},"PeriodicalIF":0.0,"publicationDate":"2025-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Cycling is an eco-friendly mode of transportation with social and environmental benefits. However, cyclist fatality rates in Europe have not shown improvement, with most fatalities occurring during car-cyclist interactions. The Communication-Oriented Advanced Driver Assistance System (CO-ADAS) introduces a light-based visual communication, projecting patterns onto the ground to indicate the vehicle's intended movement, enhancing cyclists' anticipation. Evaluating CO-ADAS in on-road experiments is constrained by safety and controllability. Virtual reality (VR) bike simulators offer a safer and controlled alternative, yet the validity of simulator-based evaluation is little discussed, especially for novel interactions like CO-ADAS, which is an unfamiliar interaction type and may elicit surprising effects. This study addresses the validation of simulator-based evaluations for an unfamiliar interaction type (CO-ADAS) and highlights the challenges of simulator validation via statistical analysis. It also emphasizes the importance of video data for action recognition and for capturing complex behavior. To this end, cyclist interactions with the CO-ADAS projection were compared between on-road and VR experiments across three levels: trajectory and speed measurements, cyclists’ reactions, and perception of CO-ADAS. Comparison results showed that while average lateral position differed between experiments, variations in lateral position and average speed showed no significant differences. Reaction comparison via video observation showed higher safety reactions in the on-road experiment, yet both experiments exhibited a consistent increase in safety reactions with CO-ADAS compared to reverse lights alone. Cyclists’ perceptions were also positive across both experiments. Overall, the results support simulator-based evaluation for assessing the safety impact of novel interactions like CO-ADAS on cyclists.
{"title":"Validation of a bike simulator for ADAS impact analysis: Road projection communication system with reversing scenario case study","authors":"Meysam Imanipour, Bertrand Barbedette, Sébastien Saudrais","doi":"10.1016/j.jcmr.2025.100104","DOIUrl":"10.1016/j.jcmr.2025.100104","url":null,"abstract":"<div><div>Cycling is an eco-friendly mode of transportation with social and environmental benefits. However, cyclist fatality rates in Europe have not shown improvement, with most fatalities occurring during car-cyclist interactions. The Communication-Oriented Advanced Driver Assistance System (CO-ADAS) introduces a light-based visual communication, projecting patterns onto the ground to indicate the vehicle's intended movement, enhancing cyclists' anticipation. Evaluating CO-ADAS in on-road experiments is constrained by safety and controllability. Virtual reality (VR) bike simulators offer a safer and controlled alternative, yet the validity of simulator-based evaluation is little discussed, especially for novel interactions like CO-ADAS, which is an unfamiliar interaction type and may elicit surprising effects. This study addresses the validation of simulator-based evaluations for an unfamiliar interaction type (CO-ADAS) and highlights the challenges of simulator validation via statistical analysis. It also emphasizes the importance of video data for action recognition and for capturing complex behavior. To this end, cyclist interactions with the CO-ADAS projection were compared between on-road and VR experiments across three levels: trajectory and speed measurements, cyclists’ reactions, and perception of CO-ADAS. Comparison results showed that while average lateral position differed between experiments, variations in lateral position and average speed showed no significant differences. Reaction comparison via video observation showed higher safety reactions in the on-road experiment, yet both experiments exhibited a consistent increase in safety reactions with CO-ADAS compared to reverse lights alone. Cyclists’ perceptions were also positive across both experiments. Overall, the results support simulator-based evaluation for assessing the safety impact of novel interactions like CO-ADAS on cyclists.</div></div>","PeriodicalId":100771,"journal":{"name":"Journal of Cycling and Micromobility Research","volume":"7 ","pages":"Article 100104"},"PeriodicalIF":0.0,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926809","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-16DOI: 10.1016/j.jcmr.2025.100103
Arnor Elvarsson , David Zani , Bryan T. Adey
Timely development of cycling infrastructure is essential to achieving societal goals such as decarbonisation and cyclist traffic safety. However, delays in infrastructure project completion persist across many planning contexts, partly due to the infrastructure planning processes. This paper addresses the lack of academic research on infrastructure planning process improvement, specifically for cycling infrastructure, by applying a structured, three-step methodology—process mapping, process analysis and improvement proposal—to the case of Canton Zürich, Switzerland. The paper includes mapping the existing cycling infrastructure planning process, identifying process-related challenges using three decision-making criteria (technical readiness, societal consensus, and political-financial prioritisation), and proposing targeted improvements. Key findings highlight the need for timely planning mandates, early-stage cost overviews, and systematic treatment of uncertainty to enhance planning process efficiency. It is argued that these process modifications can accelerate the realisation of cycling infrastructure projects and improve alignment with long-term strategic goals such as achieving net-zero carbon emissions by 2050. By bridging the gap between planning process design and infrastructure outcomes, this study contributes an approach for analysing and improving planning processes. The findings are relevant for infrastructure planners, policymakers, and researchers seeking to support more effective and efficient cycling infrastructure development.
{"title":"Fast-lane for planning cycling infrastructure: On the effectiveness and efficiency of cycling infrastructure planning processes","authors":"Arnor Elvarsson , David Zani , Bryan T. Adey","doi":"10.1016/j.jcmr.2025.100103","DOIUrl":"10.1016/j.jcmr.2025.100103","url":null,"abstract":"<div><div>Timely development of cycling infrastructure is essential to achieving societal goals such as decarbonisation and cyclist traffic safety. However, delays in infrastructure project completion persist across many planning contexts, partly due to the infrastructure planning processes. This paper addresses the lack of academic research on infrastructure planning process improvement, specifically for cycling infrastructure, by applying a structured, three-step methodology—process mapping, process analysis and improvement proposal—to the case of Canton Zürich, Switzerland. The paper includes mapping the existing cycling infrastructure planning process, identifying process-related challenges using three decision-making criteria (technical readiness, societal consensus, and political-financial prioritisation), and proposing targeted improvements. Key findings highlight the need for timely planning mandates, early-stage cost overviews, and systematic treatment of uncertainty to enhance planning process efficiency. It is argued that these process modifications can accelerate the realisation of cycling infrastructure projects and improve alignment with long-term strategic goals such as achieving net-zero carbon emissions by 2050. By bridging the gap between planning process design and infrastructure outcomes, this study contributes an approach for analysing and improving planning processes. The findings are relevant for infrastructure planners, policymakers, and researchers seeking to support more effective and efficient cycling infrastructure development.</div></div>","PeriodicalId":100771,"journal":{"name":"Journal of Cycling and Micromobility Research","volume":"7 ","pages":"Article 100103"},"PeriodicalIF":0.0,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145926812","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-13DOI: 10.1016/j.jcmr.2025.100098
Danielle Teh, Özlem Simsekoglu
Shared micro-mobility services (SMSs), such as shared bikes and shared e-scooters, have emerged as a viable, sustainable transportation option that contribute towards the transition of a more sustainable mobility ecosystem. Both digital and transportation technologies play crucial roles in ensuring the success and adoption of SMSs. Therefore, this study applied the Unified Theory of Acceptance and Use of Technology 2 (UTAUT 2) to investigate the actual use behaviour of SMSs in Norway. In addition, the study framework included intrinsic motivation from the Self-Determination Theory (SDT) as an exploratory exogenous variable. Age, gender and education level served as control variables. A total of 208 valid responses from both users and non-users of SMSs were gathered via online survey. Results were analysed using the Partial Least Squares Structural Equation Modelling (PLS-SEM) technique. The modelling results indicated that facilitating conditions, social influence, performance expectancy, habit, and intrinsic motivation positively influenced SMS use intention, which subsequently predicted the actual use behaviour of SMSs. Moreover, age was a meaningful control variable while gender and education level did not show any significance. The study also revealed that most users utilised SMSs less than 1 day per week and primarily used them on leisure trips, while the most cited reason for non-use was the lack of safety features. Based on these discoveries, practical recommendations such as improving road infrastructure, enhancing micro-mobility on-road visibility, and further optimising the mobile application may encourage a greater adoption and higher utilisation of SMSs among Norwegians.
{"title":"Factors steering use behaviour of shared micro-mobility services (SMSs) in Norway: Extended UTAUT 2 with intrinsic motivation","authors":"Danielle Teh, Özlem Simsekoglu","doi":"10.1016/j.jcmr.2025.100098","DOIUrl":"10.1016/j.jcmr.2025.100098","url":null,"abstract":"<div><div>Shared micro-mobility services (SMSs), such as shared bikes and shared e-scooters, have emerged as a viable, sustainable transportation option that contribute towards the transition of a more sustainable mobility ecosystem. Both digital and transportation technologies play crucial roles in ensuring the success and adoption of SMSs. Therefore, this study applied the Unified Theory of Acceptance and Use of Technology 2 (UTAUT 2) to investigate the actual use behaviour of SMSs in Norway. In addition, the study framework included intrinsic motivation from the Self-Determination Theory (SDT) as an exploratory exogenous variable. Age, gender and education level served as control variables. A total of 208 valid responses from both users and non-users of SMSs were gathered via online survey. Results were analysed using the Partial Least Squares Structural Equation Modelling (PLS-SEM) technique. The modelling results indicated that facilitating conditions, social influence, performance expectancy, habit, and intrinsic motivation positively influenced SMS use intention, which subsequently predicted the actual use behaviour of SMSs. Moreover, age was a meaningful control variable while gender and education level did not show any significance. The study also revealed that most users utilised SMSs less than 1 day per week and primarily used them on leisure trips, while the most cited reason for non-use was the lack of safety features. Based on these discoveries, practical recommendations such as improving road infrastructure, enhancing micro-mobility on-road visibility, and further optimising the mobile application may encourage a greater adoption and higher utilisation of SMSs among Norwegians.</div></div>","PeriodicalId":100771,"journal":{"name":"Journal of Cycling and Micromobility Research","volume":"7 ","pages":"Article 100098"},"PeriodicalIF":0.0,"publicationDate":"2025-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-11DOI: 10.1016/j.jcmr.2025.100101
Siri Hegna Berge, Vibeke Milch Uhlving, Aslak Fyhri
This study investigates accident risks associated with e-scooter use among Norwegian teenagers, focusing on trends over time, gender differences, and the role of e-scooter type. Using data from 21 205 participants aged 12–22 across two survey sets (2021/2022 and 2023), we examined the influence of gender, age, e-scooter ownership, and rental availability on accident risks, applying Generalized Linear Models with a negative binomial distribution. The findings reveal a 22 % increase in the likelihood of reporting injuries and a 57 % increase in injuries requiring medical attention in 2023 compared to 2021/2022. A follow-up survey in 2024 was issued to clarify the timing of earlier accidents, aiming to distinguish between accumulation effects and actual increases in risk. The results indicated that accident accumulation effects from earlier years inflated the reported risk among older teenagers, whereas the rise among 12–15-year-olds appears genuine. Females had a higher overall risk of accidents, while males were more likely to experience severe injuries requiring medical attention. Privately owned e-scooters were consistently associated with a higher risk of accidents than shared e-scooters, with this disparity more pronounced in 2023. While the introduction of stricter regulations, such as helmet mandates, nighttime bans, and alcohol limits, showed a potential trend towards reduced injury severity for shared e-scooters, this effect was not statistically significant. In contrast, the persistent high risk among privately owned e-scooter users and older teenagers highlights the need for additional safety measures. The study’s large sample size strengthens the reliability of the findings and their potential implications for policy and research.
{"title":"Surveying e-scooter accident risk among teenagers: The case of Norway","authors":"Siri Hegna Berge, Vibeke Milch Uhlving, Aslak Fyhri","doi":"10.1016/j.jcmr.2025.100101","DOIUrl":"10.1016/j.jcmr.2025.100101","url":null,"abstract":"<div><div>This study investigates accident risks associated with e-scooter use among Norwegian teenagers, focusing on trends over time, gender differences, and the role of e-scooter type. Using data from 21 205 participants aged 12–22 across two survey sets (2021/2022 and 2023), we examined the influence of gender, age, e-scooter ownership, and rental availability on accident risks, applying Generalized Linear Models with a negative binomial distribution. The findings reveal a 22 % increase in the likelihood of reporting injuries and a 57 % increase in injuries requiring medical attention in 2023 compared to 2021/2022. A follow-up survey in 2024 was issued to clarify the timing of earlier accidents, aiming to distinguish between accumulation effects and actual increases in risk. The results indicated that accident accumulation effects from earlier years inflated the reported risk among older teenagers, whereas the rise among 12–15-year-olds appears genuine. Females had a higher overall risk of accidents, while males were more likely to experience severe injuries requiring medical attention. Privately owned e-scooters were consistently associated with a higher risk of accidents than shared e-scooters, with this disparity more pronounced in 2023. While the introduction of stricter regulations, such as helmet mandates, nighttime bans, and alcohol limits, showed a potential trend towards reduced injury severity for shared e-scooters, this effect was not statistically significant. In contrast, the persistent high risk among privately owned e-scooter users and older teenagers highlights the need for additional safety measures. The study’s large sample size strengthens the reliability of the findings and their potential implications for policy and research.</div></div>","PeriodicalId":100771,"journal":{"name":"Journal of Cycling and Micromobility Research","volume":"7 ","pages":"Article 100101"},"PeriodicalIF":0.0,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145738837","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-11DOI: 10.1016/j.jcmr.2025.100102
James Sinclair, Jonathan Nolan
This study investigates the effect of minimum passing distance laws (also called ‘1-metre laws’ or ‘3-foot laws’) on driver behaviour. Passing distances were recorded for 70 cyclists in Perth, Western Australia before and after the introduction of a minimum passing distance law, which coincided with an advertising campaign promoting cyclist safety. The results show only a marginal reduction in the share of passes that were very close to the cyclist on roads with lower speed limits. On these roads, the closest 5% of passes were 10 cm (95% CI [2, 19] cm) further from cyclists after the law and advertising campaign. For higher speed roads, there was no effect. The very small improvement found, limited to slower roads, is considerably smaller than that of other interventions to improve cyclist safety. For instance, protected bicycle lanes have been associated with a 73 cm increase in passing distance — an effect approximately seven times greater than that of the minimum passing distance law.
{"title":"Do minimum passing distance laws for cyclists change driver behaviour?","authors":"James Sinclair, Jonathan Nolan","doi":"10.1016/j.jcmr.2025.100102","DOIUrl":"10.1016/j.jcmr.2025.100102","url":null,"abstract":"<div><div>This study investigates the effect of minimum passing distance laws (also called ‘1-metre laws’ or ‘3-foot laws’) on driver behaviour. Passing distances were recorded for 70 cyclists in Perth, Western Australia before and after the introduction of a minimum passing distance law, which coincided with an advertising campaign promoting cyclist safety. The results show only a marginal reduction in the share of passes that were very close to the cyclist on roads with lower speed limits. On these roads, the closest 5% of passes were 10 cm (95% CI [2, 19] cm) further from cyclists after the law and advertising campaign. For higher speed roads, there was no effect. The very small improvement found, limited to slower roads, is considerably smaller than that of other interventions to improve cyclist safety. For instance, protected bicycle lanes have been associated with a 73 cm increase in passing distance — an effect approximately seven times greater than that of the minimum passing distance law.</div></div>","PeriodicalId":100771,"journal":{"name":"Journal of Cycling and Micromobility Research","volume":"7 ","pages":"Article 100102"},"PeriodicalIF":0.0,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791904","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-01DOI: 10.1016/j.jcmr.2025.100086
Mohit Gupta , Debjit Bhowmick , Meead Saberi , Shirui Pan , Ben Beck
Accurate bicycling volume estimation is crucial for making informed decisions and planning about future investments in bicycling infrastructure. However, traditional link-level volume estimation models are effective for motorized traffic but face significant challenges when applied to the bicycling context because of sparse data and the intricate nature of bicycling mobility patterns. To the best of our knowledge, we present the first study to utilize a Graph Convolutional Network (GCN) architecture to model link-level bicycling volumes and systematically investigate the impact of varying levels of data sparsity (0%–99%) on model performance, simulating real-world scenarios. We have leveraged Strava Metro data as the primary source of bicycling counts across 15,933 road segments/links in the City of Melbourne, Australia. To evaluate the effectiveness of the GCN model, we benchmark it against traditional machine learning models – linear regression, support vector machines, and random forest and deep learning models – multilayer perceptron (MLP) and convolutional neural network (CNN). Our results show that the GCN model outperforms these traditional models in predicting Annual Average Daily Bicycle (AADB) counts, demonstrating its ability to capture the spatial dependencies inherent in bicycle traffic network. While GCN remains robust up to 80% sparsity, its performance declines sharply beyond this threshold, highlighting the challenges of extreme data sparsity. These findings underscore the potential of GCNs in enhancing bicycling volume estimation, while also emphasizing the need for further research on methods to improve model resilience under high-sparsity conditions. Our findings offer valuable insights for city planners aiming to improve bicycling infrastructure and promote sustainable transportation.
{"title":"Evaluating the effects of data sparsity on the link-level bicycling volume estimation: A Graph Convolutional Neural Network approach","authors":"Mohit Gupta , Debjit Bhowmick , Meead Saberi , Shirui Pan , Ben Beck","doi":"10.1016/j.jcmr.2025.100086","DOIUrl":"10.1016/j.jcmr.2025.100086","url":null,"abstract":"<div><div>Accurate bicycling volume estimation is crucial for making informed decisions and planning about future investments in bicycling infrastructure. However, traditional link-level volume estimation models are effective for motorized traffic but face significant challenges when applied to the bicycling context because of sparse data and the intricate nature of bicycling mobility patterns. To the best of our knowledge, we present the first study to utilize a Graph Convolutional Network (GCN) architecture to model link-level bicycling volumes and systematically investigate the impact of varying levels of data sparsity (0%–99%) on model performance, simulating real-world scenarios. We have leveraged Strava Metro data as the primary source of bicycling counts across 15,933 road segments/links in the City of Melbourne, Australia. To evaluate the effectiveness of the GCN model, we benchmark it against traditional machine learning models – linear regression, support vector machines, and random forest and deep learning models – multilayer perceptron (MLP) and convolutional neural network (CNN). Our results show that the GCN model outperforms these traditional models in predicting Annual Average Daily Bicycle (AADB) counts, demonstrating its ability to capture the spatial dependencies inherent in bicycle traffic network. While GCN remains robust up to 80% sparsity, its performance declines sharply beyond this threshold, highlighting the challenges of extreme data sparsity. These findings underscore the potential of GCNs in enhancing bicycling volume estimation, while also emphasizing the need for further research on methods to improve model resilience under high-sparsity conditions. Our findings offer valuable insights for city planners aiming to improve bicycling infrastructure and promote sustainable transportation.</div></div>","PeriodicalId":100771,"journal":{"name":"Journal of Cycling and Micromobility Research","volume":"6 ","pages":"Article 100086"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145624104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-23DOI: 10.1016/j.jcmr.2025.100097
Chris Larkin , Craig Robson , Alistair Ford
Low Traffic Neighbourhoods (LTNs) form an important component of urban active travel infrastructure and policy in the United Kingdom. These zones aim to prioritise street space for cycling and walking by reducing motorised traffic within the neighbourhood. While the impacts of these zones have been studied at a local scale, research often relies on bespoke datasets provided by local councils, which are frequently unavailable or do not exist for many urban areas. To examine LTNs at a larger scale, there is a need for tools and methods to identify their locations. This study develops an open, data-driven approach to identify and evaluate plausible LTNs for any Local Authority District in the UK, creating the foundations for a national LTN dataset. The methods are applied to Newcastle Upon Tyne to demonstrate the tool. First, we separate the city into neighbourhoods based on areas where people can cycle or walk comfortably before encountering a severance, such as major roads. For each neighbourhood, three metrics are produced using OpenStreetMap data to measure the transport characteristics of the zone: the density of modal filtering, the presence of traffic through-routes, and permeability difference between active modes and vehicle modes passing through a neighbourhood. Across Newcastle, we identify 215 unique neighbourhoods with 339 modal filters. Notably, 55% of neighbourhoods contain no modal filtering, while 66% have traffic through-routes. Permeability differences with neighbourhoods range from 0 m to 4046 m. Metrics are combined to provide an overall LTN plausibility score, which is visualised through automatically generated web map outputs.
{"title":"Identification of plausible low traffic neighbourhoods using open data","authors":"Chris Larkin , Craig Robson , Alistair Ford","doi":"10.1016/j.jcmr.2025.100097","DOIUrl":"10.1016/j.jcmr.2025.100097","url":null,"abstract":"<div><div>Low Traffic Neighbourhoods (LTNs) form an important component of urban active travel infrastructure and policy in the United Kingdom. These zones aim to prioritise street space for cycling and walking by reducing motorised traffic within the neighbourhood. While the impacts of these zones have been studied at a local scale, research often relies on bespoke datasets provided by local councils, which are frequently unavailable or do not exist for many urban areas. To examine LTNs at a larger scale, there is a need for tools and methods to identify their locations. This study develops an open, data-driven approach to identify and evaluate plausible LTNs for any Local Authority District in the UK, creating the foundations for a national LTN dataset. The methods are applied to Newcastle Upon Tyne to demonstrate the tool. First, we separate the city into neighbourhoods based on areas where people can cycle or walk comfortably before encountering a severance, such as major roads. For each neighbourhood, three metrics are produced using OpenStreetMap data to measure the transport characteristics of the zone: the density of modal filtering, the presence of traffic through-routes, and permeability difference between active modes and vehicle modes passing through a neighbourhood. Across Newcastle, we identify 215 unique neighbourhoods with 339 modal filters. Notably, 55% of neighbourhoods contain no modal filtering, while 66% have traffic through-routes. Permeability differences with neighbourhoods range from 0 m to 4046 m. Metrics are combined to provide an overall LTN plausibility score, which is visualised through automatically generated web map outputs.</div></div>","PeriodicalId":100771,"journal":{"name":"Journal of Cycling and Micromobility Research","volume":"6 ","pages":"Article 100097"},"PeriodicalIF":0.0,"publicationDate":"2025-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145416397","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Promoting active mobility can enhance health, reduce greenhouse gas emissions, and create more liveable cities. The choice to use active mobility for a trip, namely whether to walk or cycle instead of other modes, is known to be influenced by socioeconomic and material factors such as age, trip purpose, gradient, infrastructure, and many others. Individual attitudes and regional mobility cultures also impact this decision, but are under-researched due to data limitations. To address this, we analyse national and regional differences in active mobility in Denmark and Germany using a joint trip-level mode choice model based on large-scale survey data. Our findings show that Danes have a higher base utility for cycling compared to Germans, even after accounting for differences in socioeconomics, bicycle ownership, infrastructure, and other material factors. Only Copenhagen City, Copenhagen surroundings, and Hamburg significantly deviate from model predictions regarding active mode use, suggesting that distinct regional mobility cultures not captured by material or socioeconomic factors are at play there. These results highlight the importance of cultural factors in shaping active mode choices, supporting the need for policies addressing both cultural and infrastructural aspects to promote active mobility, while at the same time demonstrating that few regions exhibit active mobility cultures that cannot be captured using socioeconomic and material factors.
{"title":"Is it cycling culture? Explaining regional bicycle use in Denmark and Germany using a joint trip-level mode choice model","authors":"Leonard Arning , Mads Paulsen , Jeppe Rich , Heather Kaths","doi":"10.1016/j.jcmr.2025.100096","DOIUrl":"10.1016/j.jcmr.2025.100096","url":null,"abstract":"<div><div>Promoting active mobility can enhance health, reduce greenhouse gas emissions, and create more liveable cities. The choice to use active mobility for a trip, namely whether to walk or cycle instead of other modes, is known to be influenced by socioeconomic and material factors such as age, trip purpose, gradient, infrastructure, and many others. Individual attitudes and regional mobility cultures also impact this decision, but are under-researched due to data limitations. To address this, we analyse national and regional differences in active mobility in Denmark and Germany using a joint trip-level mode choice model based on large-scale survey data. Our findings show that Danes have a higher base utility for cycling compared to Germans, even after accounting for differences in socioeconomics, bicycle ownership, infrastructure, and other material factors. Only Copenhagen City, Copenhagen surroundings, and Hamburg significantly deviate from model predictions regarding active mode use, suggesting that distinct regional mobility cultures not captured by material or socioeconomic factors are at play there. These results highlight the importance of cultural factors in shaping active mode choices, supporting the need for policies addressing both cultural and infrastructural aspects to promote active mobility, while at the same time demonstrating that few regions exhibit active mobility cultures that cannot be captured using socioeconomic and material factors.</div></div>","PeriodicalId":100771,"journal":{"name":"Journal of Cycling and Micromobility Research","volume":"6 ","pages":"Article 100096"},"PeriodicalIF":0.0,"publicationDate":"2025-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145320667","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号