Bicycle Counts Research Articles

Pump or pedal? The impact of fuel prices on cycling in Germany

We analyze the relationship between fuel prices and cycling in Germany. The estimated cross-elasticity is positive and significant for utilitarian traffic, and increased in 2022 when prices became more salient. To derive our results, we exploit fuel-price variations caused by the Russian invasion of Ukraine, among other factors. Cross-elasticities are then estimated by running both OLS and IV regression models on hourly bicycle counts from 72 automated counting stations. While cross-elasticities have been estimated for low-cycling countries such as the U.S. or Australia, our results provide valuable insights for transport planning and sustainable policy making in higher-cycling countries.

Revisiting COVID-19’s Impact on Cycling: An Examination of Bicycle Count Data in Montréal

During the earlier stages of the COVID-19 pandemic, numerous studies used bicycle count data to understand the pandemic’s impact on urban cycling. However, given the context and timeliness of such studies, correcting for periods of missing data and ensuring data accuracy were not always possible. In this paper, we examine the quality of bicycle count data, using Montreal as a case study, and propose methodological improvements. By employing a missing value imputation technique and excluding sensors potentially affected by exogenous factors, we hope to provide cities with a more precise assessment of cycling trends.

More afraid of the virus than of bad weather? Exploring the link between weather conditions and cycling volume in German cities before and during the COVID-19 pandemic

Cycling can only unfold its full potential for environmental sustainability, livable cities, and health if people cycle under most seasonal and weather conditions. During the COVID-19 pandemic, many people preferred cycling over public transit, in order to avoid infection. This change may also have affected the weather–cycling association: the pandemic mainly attracted new, fair-weather cyclists; and it may have encouraged cyclists to continue cycling, even under less favorable conditions. In this paper, we explore how the COVID-19 pandemic affected the association between weather and cycling. We analyze data from bicycle counting stations in Germany, combined with city- and hour-specific weather information and a continuous measure of pandemic intensity. The data set contains more than 2.2 million hourly bicycle count observations from 69 counting stations, covering the years 2017–2021. Results from the most transit-oriented cities in the sample, Berlin and Munich suggest that the cycling volume was more dependent on weather conditions at times of high pandemic intensity. This suggests that the pandemic put new people in the saddle, but these additional cyclists were largely fair-weather cyclists who only cycled when the weather was sufficiently nice. For policy, this could set two targets. First, encourage the new pandemic-cyclists to stay on the bike and, second, support them in transitioning into year-round cyclists eventually.

Estimating daily bicycle counts with Strava data in rural and urban locations

Reliable information on daily bicycle traffic provides a fundamental basis for city planners and scientists. To estimate daily bicycle counts for various German locations with different degrees of urbanisation, we applied Generalised Boosted Regression Models. Altogether 44,136 daily datapoints from 46 counter locations covering a time period of four years were considered. Crowdsourced fitness tracker data from Strava, socio-demographics, land use and weather data were used as independent variables. Our results indicate that weather has the strongest influence on estimated bicycle counts, exceeding the relevance of fitness tracker data. In an overall model daily bicycle counts were estimated with a mean absolute percentage error (MAPE) of 27.9 %. In terms of location-specific estimations, a MAPE of 11.2 % was reached. With our approach, high-quality out-of-sample predictions are also feasible. Based on our estimations, we assume the volatility of fitness tracker user share to have a major impact on model accuracy.

Integrating link count data for enhanced estimation of deterrence functions: A case study of short-term bicycle network interventions

Owing to the lack of demand-related data from interim timespans between consecutive household surveys, it is difficult to assess short-term interventions in the bicycling network and their impact on potential accessibility. In this study, we propose a novel method that incorporates easier and cheaper demand-related data for estimating the parameters of deterrence functions of gravity-based distribution models. In particular, an origin-destination (OD) matrix-updating procedure that uses bicycle count data was incorporated into a two-step line-search algorithm for estimating a two-parameter deterrence function. The method was applied to a case study in the city of São Paulo, where potential accessibility was calculated during the years 2014–2017, which was a period interim to the two most recent household surveys. Three different models were used to estimate the generalized costs of the network links. This method proved useful for capturing accessibility variations over short timespans. Results show how demand is affected by the bicycling network and provide insights into how short-term assessments can benefit decision-making such as providing well-connected and continuous bicycling infrastructure positively influences accessibility. However, they fail as a unique solution apart from land-use policies. Finally, our proposed method can become a valuable tool for decision-making from an accessibility perspective in the short term.

A machine learning approach to predicting bicycle demand during the COVID-19 pandemic

Count-based bicycle demand models have traditionally focused on estimation rather than prediction and have been criticized for lacking a direct causal relationship between significant variables and the activity being modeled. Because they are not choice-based models, they are doubted for their ability to forecast well. The rise of machine learning techniques has given researchers tools to build better predictive models, and the tools to evaluate predictiveness. Extensive previous work in the statistics and machine learning field has shown that the best predictive model is not synonymous with the most true (or explanatory) model. The non-motorized demand modeling community could leverage these lessons learned to develop better count-based predictive models.The rise of the COVID-19 pandemic has clearly affected travel patterns, and the broad data collection has opened-up an opportunity to leverage machine learning techniques to build a predictive bicycle demand model. This study uses bicycle count data, COVID-19 data, and weather data to develop a LASSO regression model for three facilities in Austin, TX. The COVID-19 variables included both state- and local-level data between March 15, 2020, and January 31, 2021. The final model selects six variables out of 28 variables and reveals that the increase of statewide COVID-19 fatalities, statewide molecular positivity rate, and local precipitation cause a decrease in bike ridership, meanwhile maximum temperature causes an increase. The LASSO model also has a lower prediction MSE during cross validation compared to the full model. This paper aims to bring to light that our present-day demand and volume forecasting efforts would benefit tremendously from a predictive modeling approach rather than valuing the most explanatory models as the only strong forecasters of demand. In the end, modelers can use this approach to improve the forecasting ability of any count-based bicycle demand model.

Analyzing the Behavior and Growth of Cycling in Four North American Cities Before, During, and After the COVID-19 Pandemic

The paper highlights the changes in cycling patterns and ridership trends across 12 years (including the COVID-19 pandemic) in Montreal, Vancouver, Ottawa, and New York. Using data from 17 bicycle counting stations, changes in the dynamics of daily and weekly profiles before and during pandemic were determined. Additionally, the ridership demand evolution across the years was explored using models that controlled for variations in the weather. All the studied bicycle facilities experienced changes in the daily and hourly patterns in 2020 (the first year of the pandemic), tending toward recreational purposes. Significant growth in bicycle activity during the first year of the pandemic has been found, but trends for the following years (2021 and 2022) have not been studied. This study found that all counting sites located on cycling facilities primarily used for utilitarian purposes experienced a growth in ridership during 2020. Ridership on utilitarian corridors in Montreal and New York City grew considerably during the pandemic before stabilizing in 2021 and 2022. The same counting sites rapidly reverted to utilitarian hourly and daily patterns in 2021. The mixed-utilitarian bicycle facilities in Ottawa and Montreal shifted toward more recreational uses during the pandemic, though ridership did not grow in 2021 and 2022. All the counting sites in Vancouver shifted toward mixed use during the first year of the pandemic and did not show any clear signs of reverting to their utilitarian patterns.

Near misses and split routes: Comparing rider behavior, driver interaction, and route choice for cyclists

The built environment, cyclist behavior, and driver interactions can influence route choice and, ultimately, cyclist safety. Recent studies use crowdsourced data, such as Strava, to document route selection; however, aggregated data may not fully explain the factors underpinning route selection. Utilizing naturalistic methods, we analyze videos of recorded rides and GPS data from six cyclists representing three types of riders—commuters, recreational, and athletes—to explore route choices, behavior, and driver interactions in Tucson, Arizona. Our analysis of three route selection cases highlights how intersection design, driver interactions, pavement conditions, and type of riding lead cyclists to modify behaviors and select longer detours to avoid unsafe intersections. Additionally, our study combines Strava heatmaps and physical bicycle counts to explore the number of cyclists potentially facing similar factors influencing route choice. By studying cyclists with different riding aims and utilizing both Strava heatmaps and video recordings, researchers can determine the underlying conditions, identify route locations in need of improvements, and collaborate with practitioners to implement changes to increase cyclist safety through appropriate solutions. This analysis can help ensure that designs meet user expectations.

City-oriented and inclusive bicycle-vehicle crash frequency modeling through the integration of bicycle-sharing system and other surrogates

This research explores a city-oriented and inclusive bicycle-vehicle crash frequency model through the integration of bicycle-sharing system and other surrogates pertaining to the built environment like road network, land use developments, and socio-demographic characteristics. Bicycle-vehicle crash data from 2016 to 2019, along with the road network, land use, and socio-demographic characteristics within a 1-mile radius of 62 randomly selected Capital Bikeshare (CaBi) stations (referred to as stations), was used to estimate and validate bicycle-vehicle crash frequency models (negative binomial regression models with log-link distribution) and assess the role of explanatory variables. The selected stations are geographically and fairly distributed in low-risk, moderate-risk, high-risk, and very high-risk areas while covering 99.8% of the total number of recorded bicycle-vehicle crashes. From such an inclusive approach, it was observed that bicyclists are more likely to be involved in crashes in commercial, mixed-use, and densely populated areas. In addition, they are more often involved in crashes in areas with more six-lane roads. Also, many bicycle-vehicle crashes have occurred at intersections. The nonexistence of bicycle crossings could explain this finding. The findings are substantial, and characteristics related to the built environment like road network (including vehicle and bicycle infrastructure), land use developments, and socio-demographics are good surrogates for traffic volume and bicycle count.

Impact evaluation of a cycling promotion campaign using daily bicycle counters data: The case of Cycling May in Poland

The promotion of active and sustainable transport modes as an alternative to motorised individual transport has become a key policy priority in Europe, to reduce air pollution, road congestion, noise, traffic injuries, and the adverse health outcomes of sedentary lifestyles. Policymakers are implementing a wide range of measures to encourage this shift in attitudes and behaviours. These interventions are most valuable when targeting children and young people, as they weigh in habit formation and result in longer term benefits. More specifically, soft transport policies include measures like informational and educational campaigns, marketing techniques, personalized services and incentives. Among this type of policies, the gamification approach based on a competition mechanism and relative reward is applied to promote cycling behaviour, especially directed towards children. In this study we evaluate the Cycling May policy, a wide campaign aimed at promoting commuting to school by bicycle and implemented in several cities in Poland, and targeting schoolchildren and their families. We apply a quasi-experimental design based on observed daily bicycle counts on bicycle lanes in the targeted city of Gdansk and in a control city, over a three-years period. Estimates from a difference-in-difference panel regression show that the policy generated a 18% average increase in daily bicycle traffic. Despite the positive estimated effect during the intervention, our findings suggest that the behavioural change is not sustained after the intervention ends, consistently with findings from similar research.

Active transportation pilot program evaluation: A longitudinal assessment of bicycle facility density changes on use in Minneapolis

Cities around the world are adding bicycle facilities to enhance physically active travel to improve sustainable transport and public health outcomes. One of the most promising policy interventions is the use of targeted pilot programs that aim to build connected networks of bicycle facilities to increase bicycle use. In the United States, the federal Nonmotorized Transportation Pilot Program (NTPP) provided approximately $25 million to four communities (Columbia, MO, Marin County, CA, Minneapolis, MN, and Sheboygan County, WI) to test the impact of building a full network of bicycle facilities. This study examines the impact of the NTPP in Minneapolis to determine whether the addition of bicycle facilities and increased density of bicycle facilities are related to increased bicycle ridership over time. Secondarily, the study examines the impact of differing facility types (protected facilities like trails vs. non-protected facilities like on-street bike lanes) on bicycle use. The study finds that both the presence of and density of bicycle facilities emerged as significant independent predictors of bicycle counts and growth in counts over time. Compared to locations with no facility, sites with protected facilities had 113 more cyclists during the evening peak 2-hour count period (95% CI 16.19, 209.99; p = 0.02) and a greater rate of increase in cyclist counts over time. Over the study period, counts increased by 69% at locations with protected bikeways, by 26% at locations with on-street bike lanes, and by 10% at locations with no on-site facility.

Leveraging the spatial-temporal resolution of crowdsourced cycling data to improve the estimation of hourly bicycle volume

Traditional methods for measuring bicycle volume have proven to be challenging and costly, especially when planning for non-motorized facilities at network level. One potential enhancement to traditional bicycle volume measurement methods is to use crowdsourced cycling data in conjunction with other data. This study explored the potential of incorporating crowdsourced data in bicycle volume estimation methods to improve the spatial-temporal coverage and resolution. To account for the potential biases associated with crowdsourced data, the study used additional predictors to enhance estimations. Different probabilistic and Machine Learning models were tested, including the Negative Binomial (NB), K-Nearest Neighbors (KNN), Random Forest (RF), Support Vector Machines (SVM) and Artificial Neural Networks (ANN). Both RF and ANN model were found to have a better prediction capability compared to other models. For the RF model, the addition of crowdsourced bicycle counts from Strava, which had an average observed penetration rate of 7 percent across study sites, improved the model significantly by increasing its ability to explain variations in hourly bicycle volume from 65 percent (R2 = 0.65) to 71 percent (R2 = 0.71). A simulation study to assess the change in model performance based on different simulated penetration rates found that a penetration rate of about 40 percent would improve prediction capability to 96 percent. To demonstrate a practical way for integrating crowdsourced data in practice, a web-based tool that can be used to estimate bicycle counts was developed. The results of this study can assist planners and engineers to make informed decisions by providing them with a quick method for estimating bicycle volume over a network when they have counts from crowdsourced data sources.

Multi-use physical activity trails in an urban setting and cardiovascular disease: a difference-in-differences analysis of a natural experiment in Winnipeg, Manitoba, Canada

ObjectiveTo determine if expansion of multi-use physical activity trails in an urban centre is associated with reduced rates of cardiovascular disease (CVD).MethodsThis was a natural experiment with a difference in differences analysis using administrative health records and trail-based cycling data in Winnipeg, Canada. Prior to the intervention, each year, 314,595 (IQR: 309,044 to 319,860) persons over 30 years without CVD were in the comparison group and 37,901 residents (IQR: 37,213 to 38,488) were in the intervention group. Following the intervention, each year, 303,853 (IQR: 302,843 to 304,465) persons were in the comparison group and 35,778 (IQR: 35,551 to 36,053) in the intervention group. The natural experiment was the construction of four multi-use trails, 4-7 km in length, between 2010 and 2012. Intervention and comparison areas were based on buffers of 400 m, 800 m and 1200 m from a new multi-use trail. Bicycle counts were obtained from electromagnetic counters embedded in the trail. The primary outcome was a composite of incident CVD events: CVD-related mortality, ischemic heart disease, cerebrovascular events and congestive heart failure. The secondary outcome was a composite of incident CVD risk factors: hypertension, diabetes and dyslipidemia.ResultsBetween 2014 and 2018, 1,681,125 cyclists were recorded on the trails, which varied ~ 2.0-fold across the four trails (2358 vs 4264 counts/week in summer months). Between 2000 and 2018, there were 82,632 CVD events and 201,058 CVD risk events. In propensity score matched Poisson regression models, the incident rate ratio (IRR) was 1.06 (95% CI: 0.90 to 1.24) for CVD events and 0.95 (95%CI: 0.88 to 1.02) for CVD risk factors for areas within 400 m of a trail, relative to comparison areas. Sensitivity analyses indicated this effect was greatest among households adjacent to the trail with highest cycling counts (IRR = 0.85; 95% CI: 0.75 to 0.96).ConclusionsThe addition of multi-use trails was not associated with differences in CVD events or CVD risk factors, however the differences in CVD risk may depend on the level of trail use.Trial registrationTrial registration number: NCT04057417.

Challenges and Opportunities of Emerging Data Sources to Estimate Network-Wide Bike Counts

Emerging sources of mobile location data such as Strava and other phone-based apps may provide useful information for assessing bicycle activity on each link of a network. Despite their potential to complement traditional bike count programs, the representativeness and suitability of these emerging sources for producing bicycle volume estimates remain unclear. This study investigates the challenges and opportunities by fusing Strava data with short-term and permanent conventional count program data to produce bicycle volume estimations using clustering and nonparametric modeling. Analysis indicates that the concentration of permanent counters at high bicycle volume locations presents a significant challenge to produce network-wide daily volume estimations even though Strava data demonstrate potential in mitigating the estimation bias at lower-volume sites. Despite the contribution of Strava to develop reliable and spatially and temporally transferable bicycle volume estimations, significant challenges remain to rely on Strava counts alone to characterize network-level activities due to sampling bias and spatial representations. This study will help planners discern and assess the challenges and opportunities of using emerging data in bicycle planning.

The BITS Project – Making cycling data available and comparable on a European scale

Intelligent Transport Systems (ITS) could be the basis for new data driven approaches and mobility applications, as many examples from motorized means of traffic have shown. However, ITS can also contribute to an improvement of the bicycle infrastructure in European cities and to the creation of new data driven innovations that make cycling more comfortable, safe and smart. The EU funded project BITS (Bicycles and Intelligent Transport Systems) was initiated to identify the ITS cycling solutions with the highest potential and to make cycling data available for all interested stakeholders. Until this development stage of the BITS project, ITS for cyclists as bicycle counting, 3D camera systems to detect near accidents, bike sensors to measure air quality and smartphone applications have been implemented and evaluated in the whole North Sea Region (Netherlands, Belgium, Denmark, Germany, UK). The newly gathered cycling data will not stay inside the project team: To support the further expansion of European cycling networks, one of the main deliverables of the BITS project is an open bicycle data portal, the so called CyclingDataHub (CDH). All interested stakeholders get access to different categories of cycling datasets with an open license and can share their personal cycling data with the general public. The data team of the University of Oldenburg, Department of Business Informatics VLBA, collected and preprocessed all the data sets and published these on a Bicycle Data website. The related Key Performance Indicators (KPIs) should contribute to the improvement of the comparability of the different applications and regions. The harmonization of the different data sets is a big issue as a common understanding of how to collect and to publish cycling data is not established so far on the European scale. Dynamic and static KPIs as well as suitable graphs deliver new insights into bicycle use in Europe. The preprocessed and enriched cycling data can make an important contribution to the further uptake of cycling in the North Sea Region.

Abstract 10333: Recreational Multi Use Trails and Cardiovascular Disease: A Difference-in-Differences 18 Year Natural Experiment

Introduction: There is little experimental evidence of the impact of multi-use recreational trails that support physical activity on cardiovascular disease (CVD). Hypothesis: Neighbourhoods that added multi-use trails would experience a greater decline in CVD events and risk factors compared to neighbourhoods that did not. Methods: We used a difference in differences design to study the addition of four multi-use trails 4-7km in length on CVD-related outcomes using administrative health, census and built environment data available for all citizens 30 years of age and older from Winnipeg, Canada. A 400m buffer stratified intervention and comparison neighbourhoods. Bicycle counts were recorded via electromagnetic counters for 5 years on all trails. The primary and secondary outcomes were composite measures of incident CVD events (mortality, ischemic heart disease, cerebrovascular disease and congestive heart failure) and CVD risk factors (hypertension, diabetes and dyslipidemia) and assessed quarterly for 10 years prior to and 6 years following the intervention. Intervention and comparison areas were propensity score matched, using scores regressed from baseline measures of age, sex, socioeconomic indicators, and built environment attributes that support physical activity. Results: Between 2012 and 2018, 1,429,588 cyclists were recorded on the trails and cycling use varied ~2.0-fold across the trails. Between 2000 and 2018, there were 82,632 CVD events and 201,058 CVD risk events. During the 18-year natural experiment CVD event rates and risk factors declined ~33% in both comparison and intervention neighbourhoods. In propensity score matched regression models, the incident rate ratio was 1.06 (95% CI: 0.90 to 1.24) for CVD events and 0.92 (95% CI: 0.84 to 1.02) for CVD risk factors. Sensitivity analyses revealed greater effect sizes with increasing trail use (incident rate ratios for highest vs lowest cycling counts = 0.85; 95% CI: 0.75 to 0.96 vs 1.08; 95% CI: 0.92 to 1.27). Conclusions: The addition of recreational multi-use trails was not associated with changes in overall CVD events or risk factors in adjacent neighbourhoods, compared to distant neighbourhoods, however, the effects on CVD risk factors may be influenced by trail use.

Healthy for whom? Equity in the spatial distribution of cycling risks in Los Angeles, CA

IntroductionThe health benefits of cycling have widely been recognized, but cycling is also associated with health risks (e.g., pollution exposure, crash risk). Past studies of these competing health impacts have been limited in their treatment of social equity, rarely considering spatial variations in risk that could be highly salient for marginalized populations. This study investigates the health risks of cycling through the lens of social equity by considering variations in PM2.5 concentrations and crash risk across space and across sociodemographic groups. MethodsWe conducted this analysis in Los Angeles County, CA, which has a relatively high bicycle fatality rate and consistent non-attainment status for traffic-related PM2.5. We used publicly available data, including PM2.5 concentrations, crash locations, bicycle counts, and street network data, to derive measures of pollution exposure and crash risk. We performed descriptive, visualization, and regression analyses to assess how pollution exposure and crash risk vary across census block groups and are associated with area-level sociodemographic characteristics. ResultsWe found that the health risks of cycling are disproportionately high among marginalized populations (i.e. in block groups with lower income, lower educational attainment, and higher shares of racial/ethnic minority populations). Census block groups with worse outdoor air quality and higher crash risk are particularly likely to be home to low-income people of color. Even in these places, the health benefits of cycling could outweigh the risks; however, the net health benefits of cycling, accounting for these risks, are likely to be lower in marginalized communities. ConclusionsHealth impact assessments related to cycling should incorporate neighborhood-level data to better assess the distribution of benefits and risks across space and across population groups. Efforts to promote cycling should focus on making cycling safer and healthier, placing emphasis on communities where it is associated with the greatest health risks.

How Would We Cycle Today If We Had the Weather of Tomorrow? An Analysis of the Impact of Climate Change on Bicycle Traffic

Bicycle usage is significantly affected by weather conditions. Climate change is, therefore, expected to have an impact on the volume of bicycle traffic, which is an important factor in the planning and design of bicycle infrastructures. To predict bicycle traffic in a changed climate in the city of Berlin, this paper compares a traditional statistical approach to three machine learning models. For this purpose, a cross-validation procedure is developed that evaluates model performance on the basis of prediction accuracy. XGBoost showed the best performance and is used for the prediction of bicycle counts. Our results indicate that we can expect an overall annual increase in bicycle traffic of 1–4% in the city of Berlin due to the changes in local weather conditions caused by global climate change. The biggest changes are expected to occur in the winter season with increases of 11–14% due to rising temperatures and only slight increases in precipitation.

Bike Count Forecast Model with Multimodal Network Connectivity Measures

With the rise of cycling as a mode choice for commuting and short-distance delivery, as well as policy objectives encouraging this trend, bike count models are increasingly critical to transportation planning and investment. Studies have found that network connectivity plays a role in such models, but there remains a lack of measure for the connectivity of a link in a multimodal trip context. This study proposes a connectivity measure that captures the importance of a link in connecting the origins of cyclists and nearby subway stations, and incorporates it in a negative binomial regression model to forecast bike counts at links. Representative bike trips are generated with regard to bike-friendliness using the New York City transit trip planner and used to determine the deviation from the shortest path via the designated link. The measure is shown to improve model fitness with a significance level within 10%. Insights are also drawn for income levels, bike lanes, subway station availability, and average commute time of travelers.

Comparing Perceived and Objective Measures of Bikeability on a University Campus: A Case Study

Using a participatory approach, this multi-method case study compared bicyclists’ perceptions and physical indicators of bikeability on a college campus. Student focus groups discussed campus bikeability and mapped and graded common bicycle routes. Trained students conducted environmental audits with bicycle counts. In discussions, students expressed concerns about safe bicycle riding on campus, especially during peak times. Congestion with pedestrians, other bicyclists, and skateboarders created the potential for crashes. Five major routes were identified with map scores ranging from 21.4% to 70%. Audit scores ranged from 82.5% to 86.7%. Bicycle counts varied by time of day and ranged from 11 to 91 bicycles. Student perceptions of bikeability were poorer than objective assessments of routes. The audit tool did not differentiate enough between routes and did not capture conflict potential accurately. Specification is needed on audit tools to capture the potential for conflict between bicycling and other forms of active transport—especially in unique settings like college campuses. Campuses that have well-marked paths for pedestrians, bicyclists, and motor vehicles and that communicate the importance and ease of bicycling may see the most overall success when promoting active transport options.