Nothing Special   »   [go: up one dir, main page]
Next Article in Journal
Spatial–Temporal Changes in Air Pollutants in Four Provinces of Sumatra Island, Indonesia: Insights from Sentinel-5P Satellite Imagery
Previous Article in Journal
Sea-Level Rise and Growth Machines in Coastal Cities
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Urban Science
Volume 9
Issue 2
10.3390/urbansci9020041
urbansci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Weather-Driven Cycling: Developing a Predictive Model for Urban Bicycle Usage Based on Five Key Weather Factors

by
Nahid Falah
1,
Nadia Falah
2 and
Jaime Solis-Guzman
2,*
1
Landkreis Harburg, Schloßplatz 6, 21423 Winsen (Luhe), Germany
2
ArDiTec Research Group, Department of Architectural Constructions II, Higher Technical School of Building Engineering, Universidad de Sevilla, Av. Reina Mercedes 4-a, 41012 Seville, Spain
*
Author to whom correspondence should be addressed.
Urban Sci. 2025, 9(2), 41; https://doi.org/10.3390/urbansci9020041
Submission received: 29 November 2024 / Revised: 22 January 2025 / Accepted: 6 February 2025 / Published: 11 February 2025
Browse Figures
Figure 1
<p>Topography of Hamburg [<a href="#B98-urbansci-09-00041" class="html-bibr">98</a>].</p> "> Figure 2
<p>The current state of Hamburg’s bicycle infrastructure [<a href="#B98-urbansci-09-00041" class="html-bibr">98</a>].</p> "> Figure 3
<p>Comparison of the weather parameters (mean air temperature °C, mean humidity %, precipitation mm, daylight minute, mean wind speed kph) in 2017, 2019, 2021, and 2023 and the bicycle trip volume during the same years.</p> "> Figure 4
<p>Methodological framework illustrating data collection and model validation steps.</p> "> Figure 5
<p>Average monthly variability in the bicycle volume, monthly mean daylight hours, monthly mean precipitation, monthly mean humidity, and monthly mean air temperature in 2017.</p> "> Figure 6
<p>The significance of the relationship between the number of real cycling trips and the number of predicted cycling trips in 2017.</p> "> Figure 7
<p>Comparison of predicted vs. actual bicycle trips in 2019.</p> ">
Versions Notes

Abstract

:
Weather conditions significantly influence urban cycling, shaping both its frequency and intensity. This study develops a predictive model to evaluate the impact of five key meteorological factors, namely temperature, humidity, precipitation, wind speed, and daylight duration, on urban cycling trends. Using non-linear regression analysis, the research examines cycling data from 2017 to 2019 in Hamburg, Germany, comparing predicted values for 2019 with actual data to assess model accuracy. The statistical analyses reveal strong correlations between weather parameters and cycling activity, highlighting each factor’s unique influence. The model achieved high accuracy, with R2 values of 0.942 and 0.924 for 2017 and 2019, respectively. To further validate its robustness, the model is applied to data from 2021 and 2023—years not included in its initial development—yielding R2 values of 0.893 and 0.919. These results underscore the model’s reliability and adaptability across different timeframes. This study not only confirms the critical influence of weather on urban cycling patterns, but also provides a scalable framework for broader urban planning applications. Beyond the immediate findings, this research proposes expanding the model to incorporate urban factors, such as land use, population density, and socioeconomic conditions, offering a comprehensive tool for urban planners and policymakers to enhance sustainable transportation systems.

1. Introduction

Over the past century, the global population has quadrupled, while the number of passenger kilometers using transportation has increased a staggering 100-fold [1,2,3]. The rapid expansion of urbanization and the growing number of private vehicles have exacerbated the related social and environmental challenges, particularly in large-scale cities [4,5]. Issues such as noise pollution, traffic congestion, poor air quality, and greenhouse gas emissions now demand urgent solutions [3,5,6,7,8]. In response to these challenges, alternative modes of transportation, such as public transit, cycling, and walking, have gained increasing attention [9,10,11].
Among these alternatives, urban cycling has emerged as a sustainable and practical option. As cities face mounting environmental concerns and the need for efficient mobility solutions, cycling offers significant benefits [4,5,12]. It improves air quality, reduces traffic congestion and noise, lowers transportation costs, and promotes physical wellbeing in urban areas [5,11,13,14,15,16,17,18]. Research has consistently shown that cycling generates substantially lower emissions compared to motorized transport [19,20,21], making it a cornerstone of sustainable urban development [22]. Cycling is particularly effective for short commutes of up to 5 km [23,24], offering a viable alternative to cars in crowded urban areas [25]. It also serves as a crucial mode of transport for individuals without access to private vehicles, including adolescents, children, and the elderly [5,11,26,27,28]. Beyond its environmental advantages, cycling is cost effective, fast, energy efficient, and enjoyable, making it an increasingly popular choice for daily urban travel [29,30,31].
The adoption of cycling, however, is influenced by a complex interplay of factors, ranging from natural and built environmental conditions to temporal and contextual dynamics. In Table 1, the indicators identified in previous research are summarized, providing a comprehensive view of the factors influencing cycling patterns. This study focuses on one key aspect: weather conditions. By developing a predictive model that incorporates five critical weather factors, we aim to deepen our understanding of how weather drives urban cycling behavior and to provide actionable insights for fostering cycling-friendly cities.
Table 1 outlines key factors influencing urban cycling across environmental, infrastructural, and social dimensions, forming a basis for analyzing the challenges in regard to promoting cycling in the following sections.
Among various determinants, weather conditions have been identified as key factors influencing cyclists’ comfort, safety, and overall willingness to ride (Table 2). Research indicates that weather impacts cycling behavior more significantly than factors such as topography, infrastructure, land use composition, and calendar events. Unfavorable weather conditions often lead to a notable reduction in cycling trips compared to other modes of transport, primarily because cyclists are directly exposed to weather parameters, making them more vulnerable to their effects [12,46]. Numerous studies have investigated the influence of weather parameters on travel behavior. These studies consistently demonstrate that adverse weather conditions discourage people from cycling, whereas moderate weather conditions tend to encourage it, making cycling a more appealing mode of transport in moderate weather conditions. Empirical evidence highlights that inclement weather conditions suppress cycling trips to a greater extent than other travel modes, as cyclists are directly exposed to environmental elements, rendering them more susceptible to their impacts.
Table 2 summarizes the most innovative research examining the relationship between climatic factors and cycling behavior. These studies underscore the critical role that weather plays in shaping urban cycling trends and provide a foundation for developing predictive models to promote cycling in diverse weather conditions.
Table 2 summarizes the statistical relationship between weather conditions and cycling frequency. The data reveal significant variations based on temperature and precipitation levels, emphasizing the importance of weather-resilient infrastructure for cycling. This analysis serves as a basis for the subsequent methodological framework presented in this paper.

1.1. Literature Review

The background research underscores the need for more precise weather-related studies on cycling. While previous research has established a foundational understanding of the impact of weather on cycling, it often lacks the spatial and temporal resolution necessary for detailed urban planning. Achieving the establishment of a sustainable and practical mode of transport, particularly cycling, requires continuous monitoring of cycling demand throughout the year and analyzing how different weather conditions affect the number of cyclists.
Recent studies have underscored the critical role of cycling in advancing sustainable urban mobility, with diverse approaches to understanding its integration into urban systems. Hong et al. [4] demonstrated that air pollution continues to negatively impact bike-sharing usage, even during the COVID-19 pandemic, emphasizing the need for cleaner urban environments to promote cycling adoption. Similarly, Chen et al. [11] explored the non-linear effects of spatial and temporal factors on dockless bike-sharing systems, highlighting the importance of adopting localized approaches in urban cycling policies. Lanvin et al. [62] provided an updated analytical model showcasing the non-linear effects of weather conditions on bicycle traffic, underlining the complexity of environmental influences on cycling behavior. Valentini et al. [3] argued that despite cycling’s potential for decarbonization, it remains underutilized due to gaps in governance and technological integration, such as the limited deployment of e-bikes and bike-sharing systems. Werschmöller et al. [24] identified grassroots movements as key drivers in institutionalizing cycling policies, demonstrating their influence on enhancing cycling infrastructure in German cities. Gaio and Cugurullo [45] warned of the potential marginalization of cyclists in urban mobility systems dominated by autonomous vehicles, advocating for proactive policies to maintain cycling’s role in sustainable transportation. Finally, the work by Lanvin et al. [62] highlighted the necessity of radar-style frameworks to evaluate the multi-faceted impact of cycling on sustainability. Collectively, these studies emphasize the need for integrated approaches that address social, environmental, and policy dimensions, to fully harness cycling’s potential in regard to sustainable urban transitions.
These studies provide a foundation for understanding the multi-dimensional challenges of urban cycling, which this paper seeks to address by developing a more integrated and data-driven framework for enhancing urban cycling practices. Therefore, this study focuses on developing a predictive model to monitor urban bicycle usage based on key weather factors, examining their influence on cycling patterns in the city of Hamburg, Germany. In Germany, approximately 25 million bicycle trips were made daily in 2002. Over the past 15 years, this number has increased to 28 million trips per day. Furthermore, the average distance traveled by bicycle increased by 20%, rising from 3.2 km in 2002 to 3.8 km in 2017 [71].
In Hamburg, the proportion of bicycle use grew from slightly less than 10% to 15% between 2002 and 2017 [72]. Additionally, per capita bicycle ownership in Hamburg has risen significantly over the past 15 years [73]. Despite these advancements, the percentage of cyclists in Hamburg remains lower than in other cities with comparable weather patterns, such as Amsterdam (32%), Copenhagen (29%), Montreal (18.2%), Utrecht (51%), Antwerp (28.9%), Malmo (30%), and Hangzhou (30%) [71,74]. These figures highlight the urgent need for Hamburg to evaluate the relevant conditions and implement strategies to increase cycling rates, particularly when compared to similarly situated cities. In a recent study focusing on Hamburg, an agent-based model revealed that temperature discomfort occurs 33% of the time, while wind and precipitation discomfort occur 5% of the time [69,70]. These findings reinforce the profound impact of weather conditions on urban cycling trends, which significantly influences the frequency and duration of cycling activities in Hamburg [49]. The subsequent sections will provide a comprehensive literature review in regard to the specific weather factors considered in this research. Climate change exacerbates weather conditions, presenting both challenges and opportunities for urban cycling. Unfavorable weather conditions often lead to trip rescheduling, rerouting, or even cancellation, as noted by Sabir [5,61]. Predicted changes in weather patterns, such as warmer winters and more extreme weather events, are expected to significantly alter cycling patterns.
Although there has been substantial research on weather parameters and their effects on bicycle trips, most studies have analyzed these factors individually, such as temperature, precipitation, wind, sunlight, daylight hours, or seasonal variations (Table 1 and Table 2). Each study aims to understand specific aspects of the relationship between weather and cycling globally, particularly in terms of how weather conditions impact cycling rates. For example, Nosal and Miranda-Moreno [75] analyzed urban bicycle facilities in North American cities and highlighted that temperature and humidity significantly influence cycling, often with non-linear relationships. They found that precipitation negatively affects cycling flows, with greater intensity exacerbating the impact. High temperatures, heavy rainfall, and strong winds were shown to substantially reduce cycling volumes. Similarly, studies conducted in the Netherlands revealed that recreational cycling is more sensitive to weather conditions than utilitarian cycling. A weather model by K. Thomas et al. [76], demonstrated that temperature, precipitation, and wind significantly influence cycling demand, with recreational demand being more weather sensitive.
The main weather parameters that notably affect cycling volumes include air temperature (in degrees Celsius), precipitation (in millimeters), sunlight duration (in hours per day), humidity, and wind speed (in kilometers per hour) [57,60,63,77,78]. Despite these findings, most prior studies focus on single parameters or analyze them in isolation. Addressing this gap, the present study integrates all five main weather parameters, evaluating their independent and interdependent effects on cycling volumes. For the first time, this research aims to develop a formula for predicting cycling trends based on these parameters, providing actionable insights for urban planning and sustainable transport development.
Table 3 below summarizes key studies related to these five weather parameters, highlighting their role in shaping urban cycling trends.
Table 3 demonstrates how temperature, precipitation, wind, humidity, and daylight hours impact cyclist numbers, providing a comprehensive overview of the environmental factors affecting urban cycling patterns.

1.1.1. Temperature

Temperature is one of the most significant weather parameters influencing cycling activity [35]. Moderate temperatures are generally favorable, while extreme temperatures, either hot or cold, tend to deter cyclists [68]. A study by Schmitt et al. [70] found a non-linear relationship between temperature and cycling rates, indicating that higher temperatures and lower precipitation levels correlate with increased cycling activity. This highlights the importance of temperature as a critical factor in shaping urban cycling trends.

1.1.2. Precipitation

Precipitation, including rain and snow, is among the most influential weather variables affecting cycling decisions, primarily due to safety concerns and discomfort [47,66]. Increased precipitation elevates the likelihood of accidents and makes cycling less enjoyable. Research conducted in Canadian cities found that an increase in rainy and freezing days significantly reduced the annual number of cyclists [94]. Other studies [77,78,85,95,96] further emphasize the non-linear impact of rain on bicycle volume, identifying precipitation as a primary deterrent to cycling behavior.

1.1.3. Wind Speed

Wind speed plays a crucial role in cycling activity, as it directly affects the physical effort required for cycling. Strong headwinds are especially detrimental, while tailwinds have a less pronounced negative impact [80]. A study by Böcker et al. [92] demonstrated that the average daily wind speed negatively impacts cycling duration with 90% confidence, although it does not significantly affect the number of cyclists. Additionally, a machine learning approach by Mattocks [97] showed the feasibility of integrating wind speed into predictive models for bicycle usage, underlining the importance of this parameter in forecasting cycling demand.

1.1.4. Humidity

The effect of humidity on cycling behavior varies depending on the geographic context [76]. While less studied than temperature or precipitation, high humidity can lead to discomfort and excessive perspiration, reducing the appeal of cycling [41,90]. These studies suggest that humidity, although often overlooked, is a notable factor in cycling behavior, particularly in humid climates.

1.1.5. Daylight

The daylight duration significantly influences cycling patterns. Longer daylight hours not only encourage more frequent rides, but also improve visibility and safety, fostering higher bicycle usage [78,91]. Research on bike-sharing systems has demonstrated that daylight hours play a crucial role in shaping usage patterns [63,92]. These findings underline the necessity of incorporating the daylight duration into predictive models to better understand and optimize urban cycling behavior. This study presents a novel integration of five key weather parameters, namely temperature, precipitation, humidity, wind speed, and sunlight duration, into a single non-linear predictive model. While prior research often focuses on one or two weather variables, our approach addresses their combined influence on bicycle usage. Additionally, we conduct a detailed correlation analysis to evaluate potential interdependencies among these variables, ensuring the robustness of our methodology.
The five weather parameters are selected based on their proven impact on cycling behavior in prior studies. This research integrates these parameters to develop a robust predictive model for monitoring urban cycling trends.
This study investigates both the independent and combined effects of weather parameters on cycling behavior. By employing advanced data analytics and historical datasets, a predictive model is developed to monitor and analyze fluctuations in bicycle traffic in varying weather conditions. The model’s accuracy is validated using historical data and simulations across selected years, ensuring its robustness. Ultimately, the findings provide actionable insights to address weather-related challenges, optimize urban infrastructure, and promote cycling as a sustainable mode of transportation, with a focus on Hamburg’s weather patterns.

2. Materials and Methods

This study utilizes an integrated methodology to analyze the relationship between weather parameters and cycling behavior, building on established research frameworks [11,12,62]. Prior studies have explored the impact of weather conditions on urban cycling using various analytical approaches. Schmitt et al. [70] highlighted the non-linear relationship between temperature and cycling rates, emphasizing the role of moderate weather conditions in promoting cycling. Lanvin et al. [62] demonstrated the significant influence of temperature and humidity on cycling behavior in cities like Vancouver, while studies on precipitation [94] have consistently shown its negative impact on cycling volumes, particularly during rainy conditions.
Wind speed, a critical parameter, has been analyzed in studies such as that of Böcker et al. [92], which focused on the adverse effects on cycling duration, particularly in German cities. Mattocks [97] demonstrated the integration of wind speed into predictive machine learning models, emphasizing its importance in forecasting cycling demand [11]. These studies provided the foundation for including wind speed as a variable in our model.
Daylight duration has also been identified as a significant factor in shaping cycling patterns, where longer daylight hours were positively associated with increased cycling activity [12,24].
Building on insights from diverse geographic contexts, including Germany, Canada, and New Zealand, this study incorporates temperature, precipitation, wind speed, humidity, and daylight duration into a robust predictive model. By leveraging advanced data analytics and prior research, the model captures the nuanced relationships between these climatic factors and cycling behavior. It evaluates both independent and interdependent effects on cycling trends and validates its applicability using historical datasets from Hamburg. Additionally, the use of predictive models and machine learning techniques provides actionable insights for enhancing urban cycling infrastructure across varied environmental contexts. For instance, in Berlin, predictive models estimate an annual increase of 1–4% in cycling traffic due to rising temperatures, with the most significant growth anticipated during the winter months [69].
This study utilizes data from two base years (2017 and 2019) and applies a non-linear regression model to predict cycling activity. Statistical analyses and data visualizations validate the model’s reliability and accuracy, with additional testing conducted for subsequent years (2021 and 2023). The methodology is designed to ensure robust predictions and to establish the model’s flexibility for broader urban applications. The methods and models developed for this research focus not only on the five key climatic factors, namely temperature, precipitation, wind speed, humidity, and daylight duration, but also allow for the potential integration of other variables, such as air quality, traffic conditions, and socioeconomic factors. This adaptability enhances the model’s utility for urban planners and policymakers seeking to promote sustainable transportation solutions, tailored to diverse urban contexts.
This study adopts a step-by-step methodology, which includes the following components:
  • Correlation analysis and model design: The identification of strong correlations between climatic factors and cycling trends using data from the base years 2017 and 2019, followed by the design of a predictive model;
  • Evaluation of weather impacts on cycling volume: The assessment of various weather conditions, namely temperature, precipitation, wind speed, humidity, and daylight duration, on cycling volumes using SPSS;
  • Statistical analysis and model validation: The analysis of cycling trends using statistical methods such as Pearson correlation coefficients, ANOVA tests, and R-square values. The accuracy and relevance of the model are evaluated using data from the year 2019;
  • Development and testing of predictive models: The creation of predictive models for bicycle trip volumes based on weather data, with further validation performed using independent data from 2021 and 2023. These years are intentionally excluded from the model design process to ensure unbiased accuracy testing.
By leveraging historical weather and cycling data, this research aims to develop a robust, adaptable model for predicting urban cycling trends. The integration of advanced analytical methods ensures the model’s precision and applicability, providing actionable insights to support the promotion of cycling as a sustainable urban transportation solution.

2.1. Case Study Background

In this study, the city of Hamburg is selected as a case study. Located at latitude 53°33′3.9096′′ N and longitude 9°59′37.2552′′ E, Hamburg had a population of approximately 1.8 million residents in 2017, the first year of the examination (Where Is Hamburg, Germany on Map Lat Long Coordinates; [72]). Below is a brief overview of Hamburg’s key characteristics and urban structure.

2.1.1. Modal Split in Hamburg

Between 2002 and 2017, Hamburg witnessed notable shifts in its modal split. The proportion of bicycle use increased from nearly 10% to 15%, while public transport usage rose modestly from less than 20% to 22%. Pedestrian travel remained stable at approximately 27%. In contrast, the share of private vehicle use decreased significantly, from 47% to 36%. These trends, reflect an increasing preference among Hamburg residents for bicycles and public transport over private vehicles during this period [72].

2.1.2. Elevation and Topography in Hamburg

According to the topographic map of Hamburg (Figure 1), the city has a minimum elevation of −3 m, a maximum elevation of 150 m, and an average elevation of 23 m. This relatively flat topography makes Hamburg highly suitable for cycling [98]. Given that topography is a critical factor influencing cycling behavior, Hamburg offers favorable conditions for promoting bicycle usage.
Figure 1 depicts the topography of Hamburg, highlighting its relatively flat terrain, which provides favorable conditions for promoting cycling behavior in the city.

2.1.3. Current State of Hamburg’s Bicycle Infrastructure

Hamburg’s Veloroute network consists of 14 citywide cycling routes, spanning a total of approximately 280 km, of which around 80 km have been completed to date. As part of its vision to become a bicycle-friendly city, the Hamburg Senate has announced plans to develop additional cycling paths and optimize existing routes. These initiatives aim to enhance the accessibility and usability of cycling infrastructure throughout the city. For a visual overview of Hamburg’s Veloroute network, see Figure 2.
Figure 2 illustrates the current state of Hamburg’s bicycle infrastructure, highlighting key routes and their distribution across the city’s districts.
To ensure robustness and reliability, the data were collected from Hamburg’s cycling monitoring network, which includes automatic ground counters across the city. The dataset spans four years (2017, 2019, 2021, and 2023), providing 48 monthly observations. This approach ensures sufficient sample size, addressing concerns about data reliability. Future work could explore higher frequency datasets for enhanced granularity.

2.1.4. Average Weather Conditions in Hamburg

Hamburg’s climate is characterized by comfortable summers, with partly cloudy skies, and long, cold, and often windy, winters. Annually, temperatures typically range from −1 °C to 23 °C, rarely dropping below −9 °C or exceeding 29 °C. Humidity levels remain relatively stable throughout the year, fluctuating between 67% and 90%. The windiest period features average wind speeds exceeding 11.2 miles per hour [99,100].
For a comparative analysis, Figure 3 illustrates the climatic factors (mean air temperature, mean humidity, precipitation, daylight duration, and mean wind speed) across the years 2017, 2019, 2021, and 2023, alongside the volume of bicycle trips during the same time periods.
Figure 3 illustrates the relationship between weather parameters and cycling volumes (2017–2023), showing seasonal trends, where moderate temperatures and lower precipitation correspond to a higher amount of cycling activity. This highlights the influence of environmental conditions on urban cycling patterns.

2.2. Data Insights and Analytical Framework

To predict monthly cycling trends based on varying weather conditions, this study employs a non-linear regression model. Several statistical methods are utilized, including Pearson correlation coefficients, ANOVA tests (analysis of variance), R-square values (indicating the proportion of variance in terms of the dependent variable), adjusted R-square values (corrected for multiple variables), and F rates (F statistics).
The dataset is structured as follows:
  • 2017 data: used as the foundational dataset to design the predictive model;
  • 2019 data: used to validate the model’s accuracy;
  • 2021 and 2023 data: employed to evaluate the model’s performance and reliability by comparing the predicted bicycle volumes with actual data.
Various methods are employed to validate and ensure the accuracy of the model. These steps are illustrated in Figure 4.
For 2017 and 2019, the weather data (Table 4) and the average of the number of monthly bicycle trips for the years 2017, 2019, 2021, and 2023 (Table 5) are meticulously collected across 12 categories, for each month, in Hamburg. The model applied in this study is classified as non-linear, due to its ability to capture complex relationships between the weather parameters and cycling behavior. Unlike linear models, which assume a constant rate of change between variables, non-linear models account for varying effects at different levels in terms of the independent variables. For example, temperature exhibits a diminishing or reversing effect at extreme values, which is a key characteristic modeled here. This approach ensures a more accurate representation of real-world phenomena, where interactions between variables are rarely purely linear.
To evaluate the potential interdependencies among the input variables, a correlation analysis is performed using Pearson’s correlation coefficients [11]. The results showed significant correlations between temperature and daylight duration, as well as moderate correlations between precipitation and humidity. Despite these correlations, all the variables are retained in the model to ensure a comprehensive analysis of their combined effects on cycling (see Table 4).
Table 6 provides normalized values, calculated using the Min–Max normalization method. This technique scales the data to a range typically between 0 and 1, enabling better comparability between datasets. A comparison of the total number of bicycle trips in 2017 and 2019 reveals an increase in the number of trips taken in Hamburg. Additionally, the analysis highlights a clear and logical relationship between the number of bicycle trips and the various months of the year, driven by different weather conditions.
While this method assumes stationarity, it is important to note that normalization does not imply static bicycle usage trends over time. Instead, the Min–Max standardization focuses on maintaining relative relationships among variables, facilitating a robust examination of the weather impacts on cycling behavior across different temporal scales.
Table 7 demonstrates a statistically significant relationship between the number of bicycle trips in 2017 and 2019, with a Pearson correlation coefficient of 0.909 (p < 0.01).
This strong positive correlation confirms the reliability and accuracy of the data collection process, indicating that the datasets are robust and suitable for use in predictive modeling.

3. Results

To accomplish the primary objective of this study, the weather data for the year 2017 are used as the baseline dataset. These data, presented in Table 7, were collected from various meteorological sources in Hamburg and normalized to a range of 0 to 1. The same normalization method is applied to cycling data, ensuring comparability across all the datasets.

3.1. Correlation Between Weather Parameters and Cycling Trends

Table 7 is constructed using the foundational data presented in Table 8. By analyzing the relationship between climatic factors and cycling activity, the study reveals clear trends: as the temperature and daylight hours increase, the number of cyclists also increases. This relationship is visually represented in Figure 5, where all the data points have been scaled uniformly, enabling direct comparison within a single figure.
The normalized dataset for 2017 revealed clear relationships. As shown in Figure 5, the following trends are observed:
  • Bicycle trip volumes increased significantly with rising temperatures, peaking during the warmest months (May to August). Conversely, colder months (December to February) recorded the lowest bicycle activity;
  • Longer daylight hours strongly influenced cycling behavior, with the largest trip volumes occurring during months with extended daylight (April to August). This highlights the role of visibility and longer hours in which to be active in promoting cycling;
  • Precipitation showed an inverse relationship with bicycle trips. The months with higher precipitation, such as October, saw reduced cycling activity, demonstrating the deterrent effect of adverse weather conditions;
  • While less influential than temperature and daylight, humidity and wind speed exhibited minor, but measurable, impacts. Higher humidity levels and stronger winds during the winter months were associated with a slight decline in cycling activity.
As presented in Table 8, the temperature, precipitation, daylight, humidity, and wind speed data are normalized to a scale of 0 to 1, allowing for direct comparisons. Table 7 highlights a strong positive correlation (Pearson coefficient 0.909, p < 0.01) between the bicycle trip volume in 2017 and 2019, confirming the reliability of the collected data and validating the observed patterns.
Figure 5 provides a visual summary of the normalized relationship between the weather parameters and bicycle trip volume.
The following insights can be drawn from the figure:
  • Temperature and daylight hours: These are the most significant factors influencing cycling activity. Higher temperatures and longer daylight hours correlate positively with increased bicycle trips, making them crucial for predictive modeling of cycling behavior;
  • Precipitation: Rain and snowfall are primary deterrents to cycling, leading to a notable decrease in cycling activity during months with higher rainfall levels. This highlights the importance of infrastructure to reduce the impact of adverse weather conditions;
  • Humidity and wind speed: While less impactful than temperature and daylight, these climatic factors still play a role in cycling trends, especially during extreme conditions;
  • Seasonal trends: Cycling activity peaks during spring and summer, with a significant decline in the winter months, which aligns with typical weather patterns in cities like Hamburg.
To gain a more comprehensive understanding of the relationship between the weather parameters and the number of bicycle trips, the Pearson correlation coefficients for the year 2017 are calculated. The results, as presented in Table 9, quantify the degree of correlation between each weather parameter and the volume of bicycle trips.
The key findings from Table 9 are as follows:
  • Strong Positive Correlations:
    • Daylight has the highest positive correlation with bicycle trips (r = 0.957, p < 0.01), significantly boosting cycling activity;
    • Temperature is also strongly positively correlated (r = 0.947, p < 0.01), highlighting its major role in influencing cycling patterns;
  • Negative Correlations:
    • Humidity has the weakest negative correlation (r = −0.744, p < 0.01), indicating its lesser, but still measurable, impact on cycling;
    • Precipitation shows a moderate negative correlation (r = −0.515, p < 0.05), reducing cycling activity;
    • Wind speed exhibits a noticeable negative correlation (r = −0.696, p < 0.05), reflecting the challenges posed by windy conditions.
  • Covariance Analysis:
    • Temperature and daylight together strongly influence cycling volumes, explaining a significant portion of the variance in the number of bicycle trips and reinforcing their importance for predictive modeling.
To further evaluate the impact of individual climatic factors on bicycle trip volumes in 2017, an ANOVA test and R-square analysis are conducted. In this analysis, bicycle trips are considered the dependent variable, while the weather parameters serve as the independent variables. The results, summarized in Table 10, provide insights into the relative importance of each parameter in influencing cycling behavior.
The key observations from Table 10 are as follows:
4.
R-Square Values:
  • Daylight has the highest R-square value (0.900), meaning it explains the largest proportion of variance in bicycle trip volumes;
  • Wind speed has the lowest R-square value (0.020), indicating its minimal influence on cycling trends;
5.
F-Rate Analysis:
  • Daylight has the highest F rate (40.634, p < 0.01), confirming its significant impact on the amount of bicycle trips;
  • Wind speed: has the lowest F rate (0.093, p = 0.912), indicating a negligible effect on the dependent variable;
6.
Significance Levels:
  • Temperature and daylight both have p-values less than 0.05, indicating their statistically significant impact on cycling activity;
  • Precipitation, humidity, and wind speed do not show any statistically significant effects, with p-values greater than 0.05.

3.2. Analysis of Covariance

The analysis of covariance, summarized in Table 11, evaluates the individual and combined effects of temperature and daylight duration on the volume of bicycle trips. The results highlight the following key findings:
  • Individual Impacts:
Both the temperature and daylight duration have statistically significant individual effects on cycling volumes, with p-values of 0.020 and 0.014, respectively;
The Partial Eta Squared values indicate that daylight duration (0.505) has a slightly greater influence on bicycle trips than temperature (0.471);
  • Combined Effects:
The interaction term (temperature * daylight) has a p-value greater than 0.05 (0.991), indicating that the two parameters do not collectively influence cycling volumes. This suggests that their impacts are largely independent.
The key takeaway points from Table 11 are as follows:
  • Daylight’s dominance: Daylight duration has the greatest impact on cycling volumes, with the highest R-square value (0.900) and Partial Eta Squared (0.505);
  • Temperature’s role: Temperature also significantly influences cycling volumes, although its effect (Partial Eta Squared = 0.471) is slightly less than daylight’s;
  • Minimal impact of wind speed: Wind speed has a minimal effect on cycling behavior, as shown by its low R-square value (0.020) and F rate;
  • Independent impacts: Temperature and daylight duration impact cycling independently, with no significant combined effect (p = 0.991).
These findings reinforce the importance of daylight duration and temperature in driving seasonal cycling trends and can be used to inform future predictive modeling efforts.
The analysis highlights the varying degrees of influence each weather parameter exerts on cycling trends in Hamburg. Daylight duration and temperature emerged as the most significant factors, consistently driving seasonal cycling activity, while wind speed and precipitation exhibited comparatively weaker impacts. The independent nature of temperature and daylight duration effects suggests the need for targeted interventions to maximize cycling’s potential in favorable conditions.

4. Discussion

According to the classification and understanding of different relationships between weather parameters and their respective strengths and weaknesses in influencing cycling activity, it becomes evident that predicting cycling trends over time is essential. This study addresses this need by focusing on the year 2021 and establishing Equation (1). The coefficients for this equation, as detailed in Table 12, demonstrate their high precision in predicting the number of bicycle trips.
This proposed equation serves as a practical tool for estimating the monthly volume of bicycle trips by incorporating key weather parameters. It allows for a deeper understanding of how different weather conditions impact cycling and provides a reliable framework for forecasting trends in various selected time periods.
The relationship between weather parameters and bicycle trips is modeled using Equation (1), which incorporates key weather variables, such as temperature, humidity, precipitation, daylight duration, and wind speed, along with a squared term for temperature to account for its non-linear effects. The coefficients for these variables are detailed in Table 12, demonstrating their estimated impacts, significance, and confidence intervals.
Equation (1): Monthly Prediction
BTV = (α + β1 × Tmean − β2 × Hmean + β3 × Pmean + β4 × DLmean + β5 × WSmean + γ × (Tmean)2)
  • BTV = bicycle trip volume;
  • α = intercept;
  • β1—regression of numerical variables related to mean air temperature;
  • β2—regression of numerical variables related to mean humidity;
  • β3—regression of numerical variables related to precipitation;
  • β4—regression of numerical variables related to daylight;
  • β5—regression of numerical variables related to mean wind speed;
  • γ—regression (mean monthly air temperature);2
  • DLmean—means monthly daylight duration [minute];
  • Tmean—means monthly air temperature [°C];
  • Hmean—means monthly humidity [%];
  • WSmean—means monthly wind speed [kph];
  • Pmean—means monthly precipitation [mm].
Based on the data in Table 12, the following findings are made:
  • Intercept (α): Represents the baseline volume of bicycle trips when all the climatic factors are held constant (+6130);
  • Temperature (β1) and squared term (γ): Temperature positively impacts cycling activity (+9971.9), with diminishing effects at extreme values (−317.241);
  • Humidity (β2): A small positive impact (+80), reflecting its minimal, but measurable, influence;
  • Precipitation (β3): A negative coefficient (−180) indicates a deterrent effect of rainfall on cycling activity;
  • Daylight (β4): A significant positive impact (+1358.25), confirming the role of longer daylight hours in promoting cycling;
  • Wind speed (β5): The strongest negative impact (−15,582), showing that high wind speeds deter cycling.
The established Equation (1) provides a predictive framework for estimating the monthly volume of bicycle trips (BTV) based on key weather parameters, including temperature, humidity, precipitation, daylight duration, and wind speed. Each parameter’s regression coefficient, as detailed in Table 12, quantifies its individual impact on cycling activity. Additionally, the squared term for temperature accounts for its non-linear effect on bicycle trip volumes.
To enhance the practicality of the model, Equation (2) is derived by multiplying the monthly predictions from Equation (1) by the number of days in each month. This extension enables the prediction of the average daily bicycle trip volume (BTVD). Using these equations, the number of bicycle trips in Hamburg for the year 2017 is predicted, with the results summarized in Table 13. The predicted monthly volumes align closely with the actual observed data, demonstrating the robustness and accuracy of the model.
Equation (2): Daily Prediction, to estimate daily trips, Equation (1) is multiplied by the number of days in each month:
BTVD = BTV × Day
  • BTVD = bicycle trip volume per day;
  • Day = number of days in the month.
To validate the model, the correlation between real and predicted bicycle trips in 2017 is assessed. The results, presented in Table 14, indicate a Pearson correlation coefficient of 0.971 (p < 0.01). This strong positive correlation highlights the precision of the model in capturing the relationship between weather parameters and bicycle trips, confirming its suitability for predictive purposes.
The R-square value equals 0.942, according to Table 15, indicating a high level of accuracy of the proposed model in predicting the number of bicycle trips (cycling) in the city of Hamburg. Additionally, the adjusted R-square value is positive and close to the maximum amount (one) and presents the significant impact of the weather parameters on the volume of cyclists. Figure 6 further corroborates the high accuracy of the simulation.
The predictive accuracy of the proposed model is further validated by its R-square value, which equals 0.942 (as shown in Table 15). This high value demonstrates the model’s exceptional ability to explain variations in the number of bicycle trips based on the weather parameters. Furthermore, the adjusted R-square value, which is positive and close to one, confirms the significant contribution of weather parameters to the predictive model and ensures its reliability in different scenarios.
Additionally, Figure 6 illustrates the strong agreement between real and predicted values for the number of bicycle trips in 2017, visually corroborating the model’s high level of precision. This alignment highlights the robustness of the proposed equations and their applicability for predicting cycling activity in Hamburg.
Building on the success and validation in terms of Equation (2) in predicting bicycle trips in 2017, the same equation is applied to forecast the number of bicycle trips in 2019, using weather data from that year. The results in terms of this prediction are detailed in Table 16, which provides the monthly predicted bicycle trips for 2019.
To evaluate the accuracy of this simulation, the correlation coefficient between the actual number of bicycle trips and the predicted values is calculated. As presented in Table 17, a Pearson correlation coefficient of 0.961 (p < 0.01) indicates a strong and significant relationship between the actual and predicted data. This high level of agreement further validates the robustness and reliability of the predictive model across different years and weather conditions.
Additionally, as presented in Table 18, the R-square value of 0.924 further supports the model’s accuracy in explaining variations in cycling activity. The adjusted R-square value of 0.907 reinforces the model’s reliability, considering the number of predictors and the sample size. These statistical measures indicate that the proposed model effectively captures the influence of weather parameters on the number of bicycle trips. Figure 7 visually compares the predicted and actual cycling numbers for 2019, showing their strong alignment and further corroborating the high precision of the simulation.
The non-linear regression model successfully captured the complex relationships between the weather parameters and cycling behavior, particularly through the inclusion of the squared term for temperature. This approach improved the predictive accuracy of the model, as evidenced by the high R2 values, and demonstrates the model’s robustness in addressing interdependencies among the parameters.
In summary, the results demonstrate that there is a significant relationship between weather parameters and the number of cyclists in Hamburg. The Pearson correlation coefficient indicates a strong positive correlation of 0.957 between daylight duration and cyclist numbers, while humidity exhibits a negative correlation of −0.744, highlighting their respective impacts. Based on the ANOVA test, the R-square values, and the adjusted R-square values, daylight duration, with an adjusted R-square value of 0.878, has the most substantial influence on cycling activity. In contrast, wind speed, with an adjusted R-square value of 0.197, has the least impact.
The analysis of covariance is utilized to examine the simultaneous effects of weather parameters on cycling volumes. The obtained significance value of 0.991 (p > 0.05) suggests that weather parameters do not collectively influence cycling numbers, underscoring their independent impact.
To validate the accuracy of the model, the correlation coefficient between the actual and predicted number of cyclists for 2017 and 2019 is calculated. The results yielded coefficients of 0.971 and 0.961, respectively, highlighting a strong alignment between the observed and predicted values. The corresponding R-square values, 0.942 for 2017 and 0.924 for 2019, further reinforce the high precision and reliability of the proposed model in predicting cycling volumes in Hamburg.
This study uniquely examines the individual and collective effects of five key weather components on cycling levels in Hamburg. The data from 2017 was foundational in designing the model, while weather data from 2019 validated the formula by comparing the predicted values with actual observations. To further test the model’s accuracy, weather data from 2021 and 2023 are incorporated to estimate cyclist numbers and evaluate the model’s precision.
Table 19 presents the predicted monthly bicycle trips for 2021 and 2023, compared with actual observed values. By predicting cyclist numbers for every month across these two years (24 samples in total), the model’s effectiveness and dependability are rigorously assessed. As shown in Table 20, the correlation coefficients for 2021 and 2023 are 0.945 and 0.959, respectively, confirming the strong relationships between the actual and predicted values. The corresponding R-square values, 0.893 for 2021 and 0.919 for 2023, indicate that the model explains a high proportion of the variation in cycling activity, reinforcing its accuracy and reliability.
This innovative approach demonstrates the effectiveness of the proposed model in predicting cycling volumes across multiple years, achieving high levels of accuracy and precision. By successfully validating the formula with data from 2017 and 2019 and testing its applicability with data from 2021 and 2023, this study highlights the model’s robustness and establishes a pioneering framework for understanding the impact of weather conditions on cycling trends.
The correlation analysis revealed strong relationships between certain weather parameters, particularly temperature and sunlight duration. While this may introduce collinearity concerns, our model addresses these interdependencies through non-linear methods that prioritize the overall predictive performance. The inclusion of correlated variables enables a comprehensive understanding of their joint impact on cycling behavior, which would be overlooked if analyzed independently.
The results underscore the significant impact of weather parameters on urban cycling in Hamburg, validated through the achievement of high correlation coefficients and R-square values. For instance, daylight duration emerged as the most influential parameter, while wind speed exhibited the strongest deterrent effect. By validating the model with data from 2017 and 2019 and successfully applying it to forecast cycling trends for 2021 and 2023, this research demonstrates the robustness and reliability of the approach. These findings highlight the model’s value for use in urban mobility planning and sustainable transportation strategies.
Previous research on urban cycling has often focused on isolated effects of weather parameters, such as temperature [76] or precipitation [87]. However, the combined or non-linear interactions between these parameters have been largely overlooked. This study addresses this gap by integrating five critical parameters into a unified predictive model. This integration provides a more holistic understanding of how weather influences cycling behavior, bridging gaps in earlier studies.
By integrating multiple weather parameters, this model not only validates its predictive power across multiple years, but also offers a solid foundation for exploring additional variables, such as air quality indices, real-time weather forecasts, or spatially disaggregated cycling data. Incorporating higher frequency data (e.g., daily or hourly) could further enhance the precision of temporal analyses. These advancements would empower city planners and policymakers to create actionable strategies for sustainable transportation and infrastructure design.
Future research could compare different time periods to identify long-term trends, analyze broader behavioral patterns among cyclists, and incorporate additional transport modes. Testing the model in cities with diverse climates and infrastructure or expanding datasets would enhance its generalizability and accuracy. These efforts would deepen insights into cycling dynamics and sustainable urban mobility strategies.

5. Conclusions

This research on weather-driven cycling offers a comprehensive exploration of how weather parameters influence urban bicycle usage, with a focus on the city of Hamburg. The study’s innovation lies in its holistic approach, integrating historical cycling data with five key climatic factors: temperature, humidity, precipitation, wind speed, and daylight duration. By incorporating a comprehensive set of weather parameters and addressing their interdependencies, this study contributes to the literature by providing new insights into the collective influence of weather conditions on cycling behavior. Future research could extend this approach by exploring additional factors, such as air quality and real-time weather forecasts. Unlike previous studies that typically analyze only one or two factors, this research develops a robust predictive model capable of future projections. This model equips urban planners with valuable insights for devising better strategies to improve cycling conditions across different time periods.
The key findings from this study include:
Weather conditions: They play a pivotal role in determining cycling activity. Longer daylight hours and moderate temperatures encourage cycling, while high wind speeds, heavy precipitation, and elevated humidity act as significant deterrents. This study establishes clear connections between these variables and cycling trends.
Predictive model: The formulation of a multi-parameter equation significantly enhances our understanding of how weather conditions collectively impact cycling trends. This model surpasses previous approaches by providing a more detailed and integrated analysis. It enables accurate predictions of cycling activity in diverse weather conditions and timeframes.
The flexibility of the proposed model extends its applicability to future research. While this study focuses on weather parameters, the model can be expanded to include additional indicators, such as geographic features, land use patterns, population density, and socioeconomic factors. This adaptability allows the approach to be utilized for both short-term and long-term urban planning goals, from action-oriented interventions to visionary projects. The methodology can be applied across various scales, namely buildings, neighborhoods, cities, or entire regions, providing a practical framework for predicting future trends in different contexts. Future research could build upon this study by integrating additional variables, such as air quality indices and real-time weather forecasts, to enhance the predictive power and applicability of the model. Furthermore, exploring alternative approaches with adjusted weights and priorities, based on new variables, could lead to more precise and context-specific insights.
To optimize cycling conditions further, especially in adverse weather scenarios, incorporating additional criteria into the model will yield stronger scientific findings and innovative solutions. This research sets a foundation for more comprehensive analyses that combine environmental, social, and economic factors to create adaptive and inclusive urban strategies.
Practical implications: The findings from this research have substantial implications for urban planners and policymakers. By understanding the relationship between weather conditions and cycling behavior, weather-adaptive cycling infrastructure can be designed and implemented in cities. This includes sheltered bike lanes, improved drainage systems, and better lighting for shorter daylight periods, ensuring the safety and convenience of cyclists all-year round.
Furthermore, this model can be adapted for other transportation modes and urban planning challenges, making it a versatile tool for addressing diverse needs. By leveraging this framework, the quality of urban life can be enhanced in cities, sustainable mobility can be promoted, and resilience against weather-related challenges can be ensured. The integration of weather considerations into urban design is not only practical, but essential for fostering active transportation and creating livable, future-ready cities.
This study demonstrates how a data-driven, multi-faceted approach can lead to meaningful insights and actionable strategies, serving as a template for similar research across different regions and contexts. By expanding the scope and incorporating additional factors, future studies can build on this foundation to further advance the science of urban mobility and planning.
This study has certain limitations that should be acknowledged. First, the availability of recent literature and comprehensive datasets was limited, which may have impacted the scope of the analysis. Second, the integration of multiple methodologies and non-linear predictive modeling required significant time and computational resources, making the analysis process intensive. Third, access to high-frequency weather data, such as hourly observations, was constrained, which could have further improved the model’s temporal accuracy. Finally, the study focused on a single case study (Hamburg), which may restrict the generalizability of the results to other cities with different climatic and infrastructural conditions. Addressing these limitations in future research could enhance the robustness and applicability of similar studies. While this study focuses on Hamburg due to its unique and challenging weather conditions, the proposed model offers significant potential for broader applications. Future research could apply the model to cities with varying climatic and infrastructural contexts to evaluate its adaptability and generalizability. Additionally, comparative analyses between Hamburg and other urban areas could provide valuable insights into optimizing urban cycling strategies and refining the model in regard to diverse environments.

Author Contributions

Conceptualization, N.F. (Nahid Falah), N.F. (Nadia Falah), and J.S.-G.; methodology, N.F. (Nahid Falah), N.F. (Nadia Falah); software, N.F. (Nahid Falah).; validation, N.F. (Nadia Falah) and J.S.-G.; formal analysis, N.F. (Nahid Falah), N.F. (Nadia Falah); investigation, N.F. (Nahid Falah), N.F. (Nadia Falah); resources, N.F. (Nahid Falah), N.F. (Nadia Falah), and J.S.-G.; data curation, N.F. (Nadia Falah) and J.S.-G.; writing—original draft preparation N.F. (Nahid Falah), N.F. (Nadia Falah), and J.S.-G.; writing—review and editing, N.F. (Nadia Falah) and J.S.-G.; visualization, N.F. (Nahid Falah), N.F. (Nadia Falah); supervision, J.S.-G. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data Availability Statement

The original contributions presented in this study are included in the article: further inquiries can be directed to the corresponding author.

Conflicts of Interest

The authors declare that there are no conflicts of interest.

Abbreviations

BTV bicycle trip volume;
BTVD bicycle trip volume per day;
DL mean mean monthly daylight duration;
T mean mean monthly air temperature;
H mean mean monthly humidity;
WS mean mean monthly wind speed;
P mean mean monthly precipitation;
CI confidence interval;
R2 coefficient of determination;
Adjusted R2 adjusted coefficient of determination;
ANOVA analysis of variance;
F fate F statistics;
df degrees of freedom;
Sig. significance;
SPSS Statistical Package for the Social Sciences;
Std. Deviation standard deviation;
Std. Error standard error of the estimate;
N number of observations.

References

  1. Ksenofontov, M.Y.; Milyakin, S.R. The Automobilization Process and Its Determining Factors in the Past, Present, and Future. Stud. Russ. Econ. Dev. 2018, 29, 406–414. [Google Scholar] [CrossRef]
  2. Mulrow, J.; Derrible, S. Is Slower More Sustainable? The Role of Speed in Achieving Environmental Goals. Sustain. Cities Soc. 2020, 57, 102030. [Google Scholar] [CrossRef]
  3. Valentini, D.; Wangel, J.; Holmgren, S. Representations of Urban Cycling in Sustainability Transitions Research: A Review. Eur. Transp. Res. Rev. 2023, 15, 28. [Google Scholar] [CrossRef]
  4. Hong, J.; McArthur, D.P.; Sim, J.; Kim, C.H. Did Air Pollution Continue to Affect Bike Share Usage in Seoul during the COVID-19 Pandemic? J. Transp. Health 2022, 24, 101342. [Google Scholar] [CrossRef]
  5. Dzięcielski, M.; Nikitas, A.; Radzimski, A.; Caulfield, B. Understanding the Determinants of Bike-Sharing Demand in the Context of a Medium-Sized Car-Oriented City: The Case Study of Milton Keynes, UK. Sustain. Cities Soc. 2024, 114, 105781. [Google Scholar] [CrossRef]
  6. Benintendi, R.; Gòmez, E.M.; De Mare, G.; Nesticò, A.; Balsamo, G. Energy, Environment and Sustainable Development of the Belt and Road Initiative: The Chinese Scenario and Western Contributions. Sustain. Futures 2020, 2, 100009. [Google Scholar] [CrossRef]
  7. Rakha, H.A.; Almannaa, M.; Ashqar, H.; Elhenawy, M.; Ghanem, A. Traffic State Prediction: A Traveler Equity and Multi-Modal Perspective; Virginia Tech Transportation Institute: Blacksburg, VA, USA, 2019. [Google Scholar]
  8. Almannaa, M.H.; Ashqar, H.I.; Elhenawy, M.; Masoud, M.; Rakotonirainy, A.; Rakha, H. A Comparative Analysis of E-Scooter and e-Bike Usage Patterns: Findings from the City of Austin, TX. Int. J. Sustain. Transp. 2020, 15, 571–579. [Google Scholar] [CrossRef]
  9. Li, H.; Zhang, Y.; Ding, H.; Ren, G. Effects of Dockless Bike-Sharing Systems on the Usage of the London Cycle Hire. Transp. Res. Part A Policy Pract. 2019, 130, 398–411. [Google Scholar] [CrossRef]
  10. Guo, Y.; Li, Z.; Wu, Y.; Xu, C. Exploring Unobserved Heterogeneity in Bicyclists’ Red-Light Running Behaviors at Different Crossing Facilities. Accid. Anal. Prev. 2018, 115, 118–127. [Google Scholar] [CrossRef]
  11. Chen, M.; Wang, T.; Liu, Z.; Li, Y.; Tu, M. Nonlinear and Threshold Effects of the Built Environment on Dockless Bike-Sharing. Sustainability 2024, 16, 7690. [Google Scholar] [CrossRef]
  12. Wessel, J. Using Weather Forecasts to Forecast Whether Bikes Are Used. Transp. Res. Part A Policy Pract. 2020, 138, 537–559. [Google Scholar] [CrossRef]
  13. Cerutti, P.S.; Martins, R.D.; Macke, J.; Sarate, J.A.R. “Green, but Not as Green as That”: An Analysis of a Brazilian Bike-Sharing System. J. Clean. Prod. 2019, 217, 185–193. [Google Scholar] [CrossRef]
  14. Barbour, N.; Zhang, Y.; Mannering, F. A Statistical Analysis of Bike Sharing Usage and Its Potential as an Auto-Trip Substitute. J. Transp. Health 2019, 12, 253–262. [Google Scholar] [CrossRef]
  15. Iseki, H.; Tingstrom, M. A New Approach for Bikeshed Analysis with Consideration of Topography, Street Connectivity, and Energy Consumption. Comput. Environ. Urban. Syst. 2014, 48, 166–177. [Google Scholar] [CrossRef]
  16. Lovelace, R.; Philips, I. The ‘Oil Vulnerability’of Commuter Patterns: A Case Study from Yorkshire and the Humber, UK. Geoforum 2014, 51, 169–182. [Google Scholar] [CrossRef]
  17. Otero, I.; Nieuwenhuijsen, M.J.; Rojas-Rueda, D. Health Impacts of Bike Sharing Systems in Europe. Environ. Int. 2018, 115, 387–394. [Google Scholar] [CrossRef]
  18. Zhang, Y.; Mi, Z. Environmental Benefits of Bike Sharing: A Big Data-Based Analysis. Appl. Energy 2018, 220, 296–301. [Google Scholar] [CrossRef]
  19. Pucher, J.; Dill, J.; Handy, S. Infrastructure, Programs, and Policies to Increase Bicycling: An International Review. Prev. Med. 2010, 50, S106–S125. [Google Scholar] [CrossRef] [PubMed]
  20. Reiss, S. Demand Modeling and Relocation Strategies for Free-Floating Bicycle Sharing Systems; Forschungsgesellschaft für Straßen- und Verkehrswesen (FGSV): Köln, Germany, 2017. [Google Scholar]
  21. Xu, S.J.; Chow, J.Y.J. A Longitudinal Study of Bike Infrastructure Impact on Bikesharing System Performance in New York City. Int. J. Sustain. Transp. 2020, 14, 886–902. [Google Scholar] [CrossRef]
  22. Guo, Y.; Jin, H.; Qi, Z.; Hu, Z.; Ji, H.; Wan, L. Carbonized-MOF as a Sulfur Host for Aluminum–Sulfur Batteries with Enhanced Capacity and Cycling Life. Adv. Funct. Mater. 2019, 29, 1807676. [Google Scholar] [CrossRef]
  23. Saibene, G.; Manzi, G. Bike Usage in Public Bike-Sharing: An Analysis of the “BikeMi” System in Milan; University of Milan: Milan, Italy, 2015. [Google Scholar]
  24. Werschmöller, S.; Blitz, A.; Lanzendorf, M.; Arranz-López, A. The Cycling Boom in German Cities. The Role of Grassroots Movements in Institutionalizing Cycling. Int. J. Sustain. Transp. 2024, 18, 534–545. [Google Scholar] [CrossRef]
  25. Wachotsch, U. Electric Bikes Get Things Rolling: The Environmental Impact of Pedelecs and Their Potential; Universitäts-und Landesbibliothek Sachsen-Anhalt: Halle, Germany, 2014. [Google Scholar]
  26. Ji, Y.; Fan, Y.; Ermagun, A.; Cao, X.; Wang, W.; Das, K. Public Bicycle as a Feeder Mode to Rail Transit in China: The Role of Gender, Age, Income, Trip Purpose, and Bicycle Theft Experience. Int. J. Sustain. Transp. 2017, 11, 308–317. [Google Scholar] [CrossRef]
  27. Vanparijs, J.; Panis, L.I.; Meeusen, R.; de Geus, B. Exposure Measurement in Bicycle Safety Analysis: A Review of the Literature. Accid. Anal. Prev. 2015, 84, 9–19. [Google Scholar] [CrossRef] [PubMed]
  28. Wang, Y.; Berardi, U.; Akbari, H. Comparing the Effects of Urban Heat Island Mitigation Strategies for Toronto, Canada. Energy Build. 2016, 114, 2–19. [Google Scholar] [CrossRef]
  29. Ehrgott, M.; Wang, J.Y.T.; Raith, A.; Van Houtte, C. A Bi-Objective Cyclist Route Choice Model. Transp. Res. Part A Policy Pract. 2012, 46, 652–663. [Google Scholar] [CrossRef]
  30. Ziemke, D.; Metzler, S.; Nagel, K. Bicycle Traffic and Its Interaction with Motorized Traffic in an Agent-Based Transport Simulation Framework. Future Gener. Comput. Syst. 2019, 97, 30–40. [Google Scholar] [CrossRef]
  31. Ziemke, D.; Metzler, S.; Nagel, K. Modeling Bicycle Traffic in an Agent-Based Transport Simulation. Procedia Comput. Sci. 2017, 109, 923–928. [Google Scholar] [CrossRef]
  32. Heinen, E.; Van Wee, B.; Maat, K. Commuting by Bicycle: An Overview of the Literature. Transp. Rev. 2010, 30, 59–96. [Google Scholar] [CrossRef]
  33. Rietveld, P.; Daniel, V. Determinants of Bicycle Use: Do Municipal Policies Matter? Transp. Res. Part A Policy Pract. 2004, 38, 531–550. [Google Scholar] [CrossRef]
  34. Mateo-Babiano, I.; Bean, R.; Corcoran, J.; Pojani, D. How Does Our Natural and Built Environment Affect the Use of Bicycle Sharing? Transp. Res. Part A Policy Pract. 2016, 94, 295–307. [Google Scholar] [CrossRef]
  35. An, R.; Zahnow, R.; Pojani, D.; Corcoran, J. Weather and Cycling in New York: The Case of Citibike. J. Transp. Geogr. 2019, 77, 97–112. [Google Scholar] [CrossRef]
  36. Bilss, L. What If Cycling Became a Whole Lot More Popular. Retrieved from CITYLAB. 2015. Available online: http://www.citylab.com/commute/2015/11/what-if-cycling-became-a-whole-lot-more-popular/416472 (accessed on 17 November 2015).
  37. El-Assi, W.; Mahmoud, M.S.; Habib, K.N. Effects of Built Environment and Weather on Bike Sharing Demand: A Station Level Analysis of Commercial Bike Sharing in Toronto. Transportation 2017, 44, 589–613. [Google Scholar] [CrossRef]
  38. Dill, J.; Carr, T. Bicycle Commuting and Facilities in Major US Cities: If You Build Them, Commuters Will Use Them. Transp. Res. Rec. 2003, 1828, 116–123. [Google Scholar] [CrossRef]
  39. Nelson, A.C.; Allen, D. If You Build Them, Commuters Will Use Them: Association between Bicycle Facilities and Bicycle Commuting. Transp. Res. Rec. 1997, 1578, 79–83. [Google Scholar] [CrossRef]
  40. Pucher, J.; Buehler, R. Why Canadians Cycle More than Americans: A Comparative Analysis of Bicycling Trends and Policies. Transp. Policy 2006, 13, 265–279. [Google Scholar] [CrossRef]
  41. de Chardon, C.M.; Caruso, G.; Thomas, I. Bicycle Sharing System ‘Success’ Determinants. Transp. Res. Part A Policy Pract. 2017, 100, 202–214. [Google Scholar] [CrossRef]
  42. Vandenbulcke, G.; Dujardin, C.; Thomas, I.; de Geus, B.; Degraeuwe, B.; Meeusen, R.; Panis, L.I. Cycle Commuting in Belgium: Spatial Determinants and ‘Re-Cycling’Strategies. Transp. Res. Part A Policy Pract. 2011, 45, 118–137. [Google Scholar] [CrossRef]
  43. Pucher, J.; Buehler, R.; Seinen, M. Bicycling Renaissance in North America? An Update and Re-Appraisal of Cycling Trends and Policies. Transp. Res. Part A Policy Pract. 2011, 45, 451–475. [Google Scholar] [CrossRef]
  44. Muñoz, B.; Monzon, A.; Daziano, R.A. The Increasing Role of Latent Variables in Modelling Bicycle Mode Choice. Transp. Rev. 2016, 36, 737–771. [Google Scholar] [CrossRef]
  45. Gaio, A.; Cugurullo, F. Reshaping Cyclist Mobility: Understanding the Impact of Autonomous Vehicles on Urban Bicycle Users. J. Cycl. Micromobility Res. 2024, 2, 100038. [Google Scholar] [CrossRef]
  46. Nahal, T.; Mitra, R. Facilitators and Barriers to Winter Cycling: Case Study of a Downtown University in Toronto, Canada. J. Transp. Health 2018, 10, 262–271. [Google Scholar] [CrossRef]
  47. Corcoran, J.; Li, T.; Rohde, D.; Charles-Edwards, E.; Mateo-Babiano, D. Spatio-Temporal Patterns of a Public Bicycle Sharing Program: The Effect of Weather and Calendar Events. J. Transp. Geogr. 2014, 41, 292–305. [Google Scholar] [CrossRef]
  48. Böcker, L.; Dijst, M.; Faber, J.; Helbich, M. En-Route Weather and Place Valuations for Different Transport Mode Users. J. Transp. Geogr. 2015, 47, 128–138. [Google Scholar] [CrossRef]
  49. Heaney, A.K.; Carrión, D.; Burkart, K.; Lesk, C.; Jack, D. Climate Change and Physical Activity: Estimated Impacts of Ambient Temperatures on Bikeshare Usage in New York City. Environ. Health Perspect. 2019, 127, 37002. [Google Scholar] [CrossRef] [PubMed]
  50. Smith, M.S.; Kauermann, G. Bicycle Commuting in Melbourne during the 2000s Energy Crisis: A Semiparametric Analysis of Intraday Volumes. Transp. Res. Part B Methodol. 2011, 45, 1846–1862. [Google Scholar] [CrossRef]
  51. Winters, M.; Buehler, R.; Götschi, T. Policies to Promote Active Travel: Evidence from Reviews of the Literature. Curr. Environ. Health Rep. 2017, 4, 278–285. [Google Scholar] [CrossRef] [PubMed]
  52. Borgnat, P.; Abry, P.; Flandrin, P.; Robardet, C.; Rouquier, J.-B.; Fleury, E. Shared Bicycles in a City: A Signal Processing and Data Analysis Perspective. Adv. Complex. Syst. 2011, 14, 415–438. [Google Scholar] [CrossRef]
  53. Brandenburg, C.; Matzarakis, A.; Arnberger, A. Weather and Cycling—A First Approach to the Effects of Weather Conditions on Cycling. Meteorol. Appl. J. Forecast. Pract. Appl. Train. Tech. Model. 2007, 14, 61–67. [Google Scholar] [CrossRef]
  54. Zeng, L.; Lu, J.; Li, W.; Li, Y. A Fast Approach for Large-Scale Sky View Factor Estimation Using Street View Images. Build. Environ. 2018, 135, 74–84. [Google Scholar] [CrossRef]
  55. Miranda-Moreno, L.F.; Nosal, T. Weather or Not to Cycle: Temporal Trends and Impact of Weather on Cycling in an Urban Environment. Transp. Res. Rec. 2011, 2247, 42–52. [Google Scholar] [CrossRef]
  56. Nuruzzaman, M. Urban Heat Island: Causes, Effects and Mitigation Measures-a Review. Int. J. Environ. Monit. Anal. 2015, 3, 67–73. [Google Scholar] [CrossRef]
  57. Gallop, C.; Tse, C.; Zhao, J. A Seasonal Autoregressive Model of Vancouver Bicycle Traffic Using Weather Variables. In Proceedings of the Transportation Research Board 91st Annual Meeting, Washington, DC, USA, 22–26 January 2012. [Google Scholar]
  58. Chokhachian, A.; Perini, K.; Giulini, S.; Auer, T. Urban Performance and Density: Generative Study on Interdependencies of Urban Form and Environmental Measures. Sustain. Cities Soc. 2020, 53, 101952. [Google Scholar] [CrossRef]
  59. Thomas, T.; Jaarsma, R.; Tutert, B. Exploring Temporal Fluctuations of Daily Cycling Demand on Dutch Cycle Paths: The Influence of Weather on Cycling. Transportation 2013, 40, 1–22. [Google Scholar] [CrossRef]
  60. Schmiedeskamp, P.; Zhao, W. Estimating Daily Bicycle Counts in Seattle, Washington, from Seasonal and Weather Factors. Transp. Res. Rec. 2016, 2593, 94–102. [Google Scholar] [CrossRef]
  61. Sabir, M. Weather and Travel Behaviour. Ph.D. Thesis, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands, 2011. [Google Scholar]
  62. Lanvin, A.; Michel, P.; Charléty, J.; Chasse, A. Weathering Heights: An Updated Analytical Model of the Nonlinear Effects of Weather on Bicycle Traffic. J. Cycl. Micromobility Res. 2024, 2, 100031. [Google Scholar] [CrossRef]
  63. Tin, S.T.; Woodward, A.; Robinson, E.; Ameratunga, S. Temporal, Seasonal and Weather Effects on Cycle Volume: An Ecological Study. Environ. Health 2012, 11, 12. [Google Scholar] [CrossRef] [PubMed]
  64. Oja, P.; Titze, S.; Bauman, A.; De Geus, B.; Krenn, P.; Reger-Nash, B.; Kohlberger, T. Health Benefits of Cycling: A Systematic Review. Scand. J. Med. Sci. Sports 2011, 21, 496–509. [Google Scholar] [CrossRef] [PubMed]
  65. Heinen, E.; Kamruzzaman, M.; Turrell, G. The Public Bicycle-Sharing Scheme in Brisbane, Australia: Evaluating the Influence of Its Introduction on Changes in Time Spent Cycling amongst a Middle-and Older-Age Population. J. Transp. Health 2018, 10, 56–73. [Google Scholar] [CrossRef]
  66. Gebhart, K.; Noland, R.B. The Impact of Weather Conditions on Bikeshare Trips in Washington, DC. Transportation 2014, 41, 1205–1225. [Google Scholar] [CrossRef]
  67. Bjørner, T. The Advantages of and Barriers to Being Smart in a Smart City: The Perceptions of Project Managers within a Smart City Cluster Project in Greater Copenhagen. Cities 2021, 114, 103187. [Google Scholar] [CrossRef]
  68. Nielsen, T.A.S.; Skov-Petersen, H. Bikeability–Urban Structures Supporting Cycling. Effects of Local, Urban and Regional Scale Urban Form Factors on Cycling from Home and Workplace Locations in Denmark. J. Transp. Geogr. 2018, 69, 36–44. [Google Scholar] [CrossRef]
  69. Galich, A.; Nieland, S. The Impact of Weather Conditions on Mode Choice in Different Spatial Areas. Future Transp. 2023, 3, 56. [Google Scholar] [CrossRef]
  70. Schmitt, A.U.; Burgemeister, F.; Dorff, H.; Finn, T.; Hansen, A.; Kirsch, B.; Lange, I.; Radtke, J.; Ament, F. Assessing the Weather Conditions for Urban Cyclists by Spatially Dense Measurements with an Agent-Based Approach. Meteorol. Appl. 2023, 30, e2164. [Google Scholar] [CrossRef]
  71. Nobis, C. Mobilität in Deutschland—MiD: Analysen Zum Radverkehr Und Fußverkehr; Bundesministerium für Verkehr und digitale Infrastruktur: Bonn, Germany, 2019. [Google Scholar]
  72. Metropolregion, H. Kurzreport. 2018. Available online: https://www.hamburg.de/resource/blob/193420/6a4c42b749f760f1a99c62e8b516bc9c/mid-2017-kurzreport-stadt-hamburg-infas-data.pdf (accessed on 17 November 2022).
  73. infas; DLR; IVT; Infas 360. Mobilität in Deutschland—MiD: Regionalbericht: Metropolregion Hamburg Und Hamburger Verkehrsverbund GmbH; Bundesministerium für Verkehr und digitale Infrastruktur: Bonn, Germany, 2019. [Google Scholar]
  74. Global Bicycle Cities Index 2019|Coya. Available online: https://www.coya.com/bike/index-2019 (accessed on 11 April 2021).
  75. Nosal, T.; Miranda-Moreno, L.F. The Effect of Weather on the Use of North American Bicycle Facilities: A Multi-City Analysis Using Automatic Counts. Transp. Res. Part A Policy Pract. 2014, 66, 213–225. [Google Scholar] [CrossRef]
  76. Thomas, K.; Stone, M.R.; Thompson, K.G.; Gibson, A.S.C.; Ansley, L. Reproducibility of Pacing Strategy during Simulated 20-Km Cycling Time Trials in Well-Trained Cyclists. Eur. J. Appl. Physiol. 2012, 112, 223–229. [Google Scholar] [CrossRef]
  77. Wadud, Z. Cycling in a Changed Climate. J. Transp. Geogr. 2014, 35, 12–20. [Google Scholar] [CrossRef]
  78. Rose, G.; Ahmed, F.; Figliozzi, M.; Jakob, C. Quantifying and Comparing Effects of Weather on Bicycle Demand in Melbourne, Australia, and Portland, Oregon. In Proceedings of the Transportation Research Board 90th Annual Meeting, Washington, DC, USA, 23–27 January 2011. [Google Scholar]
  79. Flynn, B.S.; Dana, G.S.; Sears, J.; Aultman-Hall, L. Weather Factor Impacts on Commuting to Work by Bicycle. Prev. Med. 2012, 54, 122–124. [Google Scholar] [CrossRef] [PubMed]
  80. Richardson, A.J. Seasonal and Weather Impacts on Urban Cycling Trips; TUTI report 1-2000; The Urban Transport Institute: Victoria, BC, Canada, 2000; Volume 1. [Google Scholar]
  81. Saneinejad, S.; Roorda, M.J.; Kennedy, C. Modelling the Impact of Weather Conditions on Active Transportation Travel Behaviour. Transp. Res. D Transp. Environ. 2012, 17, 129–137. [Google Scholar] [CrossRef]
  82. Helbich, M.; Böcker, L.; Dijst, M. Geographic Heterogeneity in Cycling under Various Weather Conditions: Evidence from Greater Rotterdam. J. Transp. Geogr. 2014, 38, 38–47. [Google Scholar] [CrossRef]
  83. Noland, R.B.; Ishaque, M.M. Smart Bicycles in an Urban Area: Evaluation of a Pilot Scheme in London. J. Public. Trans. 2006, 9, 5. [Google Scholar] [CrossRef]
  84. Phung, J.; Rose, G. Temporal Variations in Usage of Melbourne’s Bike Paths. In Proceedings of the 30th Australasian Transport Research Forum, Melbourne, Australia, 25–27 September 2007. [Google Scholar]
  85. Buehler, R.; Pucher, J. Cycling to Work in 90 Large American Cities: New Evidence on the Role of Bike Paths and Lanes. Transportation 2012, 39, 409–432. [Google Scholar] [CrossRef]
  86. Lewin, A. Temporal and Weather Impacts on Bicycle Volumes. In Proceedings of the Transportation Research Board 90th Annual Meeting, Washington, DC, USA, 23–27 January 2011. [Google Scholar]
  87. Ahmed, F.; Rose, G.; Jacob, C. Impact of Weather on Commuter Cyclist Behaviour and Implications for Climate Change Adaptation. In Proceedings of the Australasian Transport Research Forum, Canberra, Australia, 29 September–1 October 2010. [Google Scholar]
  88. Creemers, L.; Wets, G.; Cools, M. Meteorological Variation in Daily Travel Behaviour: Evidence from Revealed Preference Data from the Netherlands. Theor. Appl. Climatol. 2015, 120, 183–194. [Google Scholar] [CrossRef]
  89. Ashqar, H.I.; Elhenawy, M.; Rakha, H.A. Modeling Bike Counts in a Bike-Sharing System Considering the Effect of Weather Conditions. Case Stud. Transp. Policy 2019, 7, 261–268. [Google Scholar] [CrossRef]
  90. Behan, K.; Lea, N.S. Benchmarking Active Transportation in Canadian Cities; Clean Air Partnership: Toronto, ON, USA, 2012. [Google Scholar]
  91. Böcker, L.; Thorsson, S. Integrated Weather Effects on Cycling Shares, Frequencies, and Durations in Rotterdam, the Netherlands. Weather Clim. Soc. 2014, 6, 468–481. [Google Scholar] [CrossRef]
  92. Flynn, J.R.; Flynn, J.R. Are We Getting Smarter?: Rising IQ in the Twenty-First Century; Cambridge University Press: Cambridge, UK, 2012; ISBN 1107028094. [Google Scholar]
  93. Winters, M.; Friesen, M.C.; Koehoorn, M.; Teschke, K. Utilitarian Bicycling: A Multilevel Analysis of Climate and Personal Influences. Am. J. Prev. Med. 2007, 32, 52–58. [Google Scholar] [CrossRef] [PubMed]
  94. Ahmed, F.; Rose, G.; Jacob, C. Examining the Impact of Changes in Weather on Commuter Cyclist’s Travel Behavior. In Proceedings of the 92nd Transportation Research Board Annual Meeting, Washington, DC, USA, 13–17 January 2013. [Google Scholar]
  95. Mathisen, T.A.; Annema, J.A.; Kroesen, M. The Effects of Weather and Climate Change on Cycling in Northern Norway. Eur. J. Transp. Infrastruct. Res. 2015, 15. [Google Scholar] [CrossRef]
  96. Mattocks Machine Learning Pipeline for Predicting Bike Usage from Weather Forecasts: Part 2|by Philip Mattocks|Medium. Available online: https://medium.com/@philipmattocks/machine-learning-pipeline-for-predicting-bike-usage-from-weather-forecasts-part-2-2ed1f5e9834a (accessed on 4 August 2024).
  97. Hamburg Topographic Map, Elevation, Relief. Available online: https://en-gb.topographic-map.com/maps/r7py/Hamburg/ (accessed on 17 November 2024).
  98. Alliance for Cycling. Available online: https://www.hamburg.de/contentblob/7182614/5dfef803916523ef5784e59b7c5c4107/data/buendnis-fuer-den-radverkehr-download-english.pdf (accessed on 17 November 2024).
  99. Average Weather in Hamburg, Germany, Year Round—Weather Spark. Available online: https://weatherspark.com/y/68301/Average-Weather-in-Hamburg-Germany-Year-Round (accessed on 17 November 2024).
  100. Climate and Average Monthly Weather in Hamburg (Hansestadt Hamburg), Germany. Available online: https://en.climate-data.org/europe/germany/hamburg/hamburg-69/ (accessed on 17 November 2024).
Urbansci 09 00041 g001
Figure 1. Topography of Hamburg [98].
Figure 1. Topography of Hamburg [98].
Urbansci 09 00041 g001
Urbansci 09 00041 g002
Figure 2. The current state of Hamburg’s bicycle infrastructure [98].
Figure 2. The current state of Hamburg’s bicycle infrastructure [98].
Urbansci 09 00041 g002
Urbansci 09 00041 g003
Figure 3. Comparison of the weather parameters (mean air temperature °C, mean humidity %, precipitation mm, daylight minute, mean wind speed kph) in 2017, 2019, 2021, and 2023 and the bicycle trip volume during the same years.
Figure 3. Comparison of the weather parameters (mean air temperature °C, mean humidity %, precipitation mm, daylight minute, mean wind speed kph) in 2017, 2019, 2021, and 2023 and the bicycle trip volume during the same years.
Urbansci 09 00041 g003
Urbansci 09 00041 g004
Figure 4. Methodological framework illustrating data collection and model validation steps.
Figure 4. Methodological framework illustrating data collection and model validation steps.
Urbansci 09 00041 g004
Urbansci 09 00041 g005
Figure 5. Average monthly variability in the bicycle volume, monthly mean daylight hours, monthly mean precipitation, monthly mean humidity, and monthly mean air temperature in 2017.
Figure 5. Average monthly variability in the bicycle volume, monthly mean daylight hours, monthly mean precipitation, monthly mean humidity, and monthly mean air temperature in 2017.
Urbansci 09 00041 g005
Urbansci 09 00041 g006
Figure 6. The significance of the relationship between the number of real cycling trips and the number of predicted cycling trips in 2017.
Figure 6. The significance of the relationship between the number of real cycling trips and the number of predicted cycling trips in 2017.
Urbansci 09 00041 g006
Urbansci 09 00041 g007
Figure 7. Comparison of predicted vs. actual bicycle trips in 2019.
Figure 7. Comparison of predicted vs. actual bicycle trips in 2019.
Urbansci 09 00041 g007
Table 1. Key indicators influencing cycling patterns and relevant studies.
Table 1. Key indicators influencing cycling patterns and relevant studies.
IndicatorsResearch
Topography[15,32,33,34]
Cycling infrastructure[34,35,36,37,38,39,40,41,42,43]
Social and cultural characteristics [36,44,45]
Air pollution[4,12,35,46,47,48,49]
Economic conditions and population density[5,24,50,51]
Different periods of time (day and month)[34,47,50,52,53]
Structure of the environmental conditions[54,55,56]
Weather parameters[12,32,35,40,46,47,48,49,57,58,59,60,61,62]
Table 2. Key studies on the impact of weather on urban cycling behavior.
Table 2. Key studies on the impact of weather on urban cycling behavior.
Main FocusStudies
Insights into how weather affects cycling in North America, emphasizing immediate and lagged effects.[55,63]
Moderate temperatures encourage higher participation in cycling.[64]
Real-time weather information impacts cyclists; precipitation has immediate effects on choices.[12,65]
Long-term effects of seasonal weather variations on urban cycling participation.[12,66]
Seasonal fluctuations in urban cycling trends; importance of flexible policies and infrastructure.[67]
Higher temperatures and less precipitation linked to higher cycling rates; wind and cold reduce cycling.[68]
Predictive models show rising temperatures increase cycling, particularly during colder months.[62,69]
Agent-based model showing importance of spatially dense weather measurements compared to single stations.[11,70]
Table 3. Summary of studies on key weather parameters influencing urban cycling.
Table 3. Summary of studies on key weather parameters influencing urban cycling.
Weather ParameterMain FindingsKey Studies
TemperatureRising temperatures lead to an increase in the number of cyclists.[63,66,79]
A 0.56 °C temperature increase in the United States results in a 3% rise in the number of cyclists.[80]
A 1 degree temperature increase in Vancouver leads to a 1.65% increase in the number of cyclists.[57]
A 1 degree temperature increase in Auckland results in a 3.2% increase in the number of cyclists.[63]
The ideal cycling temperature is 25–28 °C.[81]
Most cycling trips occur within the temperature range from 26.7 °C to 31.7 °C, with temperatures above 32 °C reducing the number of cyclists.[66]
Cyclists are sensitive to temperatures below 15 °C.[63,82]
High or low temperatures can deter cycling, particularly among recreational cyclists.[40,49,53,83]
The “ideal” cycling temperature range is between 17 °C and 33 °C.[35]
PrecipitationRainfall is linked to a decrease in cycling numbers.[84]
Daily rainfall of approximately 8 mm results in a 50% reduction in the bicycle volume compared to rain-free days.[81,85]
For every 1 mm of rainfall, there is a 10.6% decrease in the number of cyclists.[63]
When daily rainfall ranges from 0.2 to 2 mm, the cyclist count decreases by 8–19%.[85]
Rainfall and temperature variables significantly reduce cycling volume in city area.[86]
WindAn increase in wind speed has a relative, adverse effect on the number of cyclists and winds exceeding 5 km/h lead to a 17% decrease in bicycle trips; also, for every 1.6 km/h increase in wind speed, there is a 5% decrease in the number of cyclists.[47,66,87]
Various wind speeds have differing impacts on cycling.[76,88,89]
Wind speeds ranging from 25 to 52 km/h result in an 11% to 23% reduction in the cyclist count.[85]
HumidityAn increase in humidity has been associated with a decrease in cycling.[41,47,90]
A one percent increase in humidity results in a 0.08 percent decrease in bicycle traffic in Vancouver.[57]
DaylightThere is a direct relationship between the hours in the day and the number of cyclists.[91,92]
Daylight effect appears to be statistically insignificant or very low.[78,93]
Table 4. Descriptive statistics for weather parameters in 2017 and 2019.
Table 4. Descriptive statistics for weather parameters in 2017 and 2019.
YearWeather ParameterUnitNMeanStd. DeviationVarianceMinimumMaximum
2017Mean air temperature°C1212.33337.1647351.3333.0022.00
Mean humidity%1283.005.5267930.54575.0091.00
Precipitationmm1281.24126.70713.32450.40133.60
Daylight durationMinute12734.3333205.0043642,026.788449.001018.00
Mean wind speedkph129.96501.617902.6188.2313.22
2019Mean air temperature°C1210.58335.8691534.4472.0019.00
Mean humidity%1278.41587.7362259.84966.1090.95
Precipitationmm1277.94027.601762.32238.47128.89
Daylight durationMinute12737.0833202.8539441,149.720450.001020.00
Mean wind speedkph128.97501.592372.5366.8912.70
Table 5. Descriptive statistics for bicycle volume parameters in 2017, 2019, 2021, and 2023.
Table 5. Descriptive statistics for bicycle volume parameters in 2017, 2019, 2021, and 2023.
Bicycle volume 2017JanFebMarAprMayJun
15,461,51915,436,52426,700,01626,690,66237,829,65838,418,756
JulAugSepOctNovDec
38,469,07139,445,52529,868,04924,295,89021,402,48914,481,840
Bicycle volume 2019JanFebMarAprMayJun
16,192,36822,668,73520,880,96135,521,64634,299,10341,290,866
JulAugSepOctNovDec
39,033,65639,853,84032,745,00025,814,83620,483,28116,804,203
Bicycle volume 2021JanFebMarAprMayJun
15,554,08414,145,60423,731,62625,850,79828,109,30240,523,525
JulAugSepOctNovDec
38,121,86834,115,33739,177,34130,394,09123,383,60615,632,669
Bicycle volume 2023JanFebMarAprMayJun
19,439,09719,767,59321,507,95527,597,12735,695,32738,823,318
JulAugSepOctNovDec
29,612,76729,065,78835,629,36921,084,47517,246,0709,094,792
Table 6. Normalization of the monthly average of the number of bicycle trips in Hamburg in 2017 and 2019.
Table 6. Normalization of the monthly average of the number of bicycle trips in Hamburg in 2017 and 2019.
Bicycle trips 2017JanFebMarAprMayJun
0.047060.04690.08120.08120.11510.1169
JulAugSepOctNovDec
0.11710.12000.09090.07390.06510.0440
Bicycle trips 2019JanFebMarAprMayJun
0.046850.06550.06040.10270.09920.1194
JulAugSepOctNovDec
0.11290.11530.09470.07460.05920.0486
Table 7. The Pearson correlation in terms of the number of bicycle trips made in Hamburg in 2017 and 2019.
Table 7. The Pearson correlation in terms of the number of bicycle trips made in Hamburg in 2017 and 2019.
CorrelationsBicycle Trips in 2017Bicycle Trips in 2019
Bicycle trip volume 2017Pearson Correlation10.909 **
Sig. (2-tailed) 0.000
N1212
Bicycle trip volume 2019Pearson Correlation0.909 **1
Sig. (2-tailed)0.000
N1212
** The correlation is significant at the 0.01 level (2-tailed).
Table 8. Normalized values in terms of the weather parameters in 2017.
Table 8. Normalized values in terms of the weather parameters in 2017.
Year 2017Number of Bicycle TripsTemperature
°C		Humidity
%		Precipitation mmDaylight
Minute		Wind
kph
MonthJanuary0.0470670290.020.0890.060.050.073
February0.0469909410.0270.0840.0610.060.069
March0.0812785880.0470.0790.0710.080.086
April0.0812501130.0810.0770.0510.0950.11
May0.1151587770.1140.0750.0840.1080.083
June0.1169520730.1350.0770.1370.1150.091
July0.1171052390.1480.0810.1120.1110.068
August0.1200777020.1410.080.0590.0990.071
September0.0909225240.0970.0860.0780.0860.073
October0.0739600920.0870.0880.1240.0710.096
November0.0651521740.0470.0910.0880.0580.076
December0.0440847490.0270.0890.0630.050.098
Table 9. Pearson correlation between the number of bicycle trips and weather parameters in 2017.
Table 9. Pearson correlation between the number of bicycle trips and weather parameters in 2017.
Number of
Bicycle Trips		Temperature
°C		Humidity
%		Precipitation mmDaylight
Minute		Wind
kph
Pearson Correlation10.947 **−0.744 **−0.515 *0.957 **−0.696 *
Sig. (2-tailed) 0.0000.0060.0470.0000.012
Sum of Squares and Cross-products0.0090.014−0.0010.0020.007−0.013
Covariance0.0010.0010.0000.0000.001−0.001
N121212121212
** The correlation is significant at the 0.01 level (2-tailed). * The correlation is significant at the 0.05 level (2-tailed).
Table 10. ANOVA test carried out in SPSS.
Table 10. ANOVA test carried out in SPSS.
Dependent Variable: Bicycle Trip Volume 2017Model SummaryANOVA
RR SquareAdjusted R SquareStd. Error of EstimateFSig.
Independent variableMean air temperature (o C)0.9450.8930.8703,445,632.51237.7420.000
Mean humidity (%)0.7210.5200.4137,314,097.3674.8750.037
Mean precipitation (mm)0.5750.3310.1828,635,688.3342.2250.164
Mean daylight duration (Minute)0.9490.9000.8783,333,413.40840.6340.000
Mean wind speed (kph)0.1420.0200.19710,449,116.2860.0930.912
Table 11. Tests of between-subject effects (dependent variable: bicycle trip volume 2017).
Table 11. Tests of between-subject effects (dependent variable: bicycle trip volume 2017).
SourceType III Sum of SquaresdfMean SquareFSig.Partial Eta Squared
Corrected Model 949,932,721,313,230.000 a3 316,644,240,437,743.300 47.714 0.0000.947
Intercept 357,175,731,059.297 1 357,175,731,059.297 0.054 0.0420.029
Temperature
(°C)		 7,164,973,489,802.059 1 7,164,973,489,802.059 1.080 0.0200.471
Daylight (Minute) 23,781,541,670,782.650 1 23,781,541,670,782.650 3.584 0.0140.505
Temperature * Daylight 953,712,204.422 1 953,712,204.422 0.000 0.9910.000
Error 53,090,519,077,594.910 8
Total 9,995,710,685,640,824.000 12
Corrected Total 1,003,023,240,390,824.900 11
a R2 = 0.947 (adjusted R2 = 0.927).
Table 12. Investigating the relationship between the number of trips made by bicycle in the 12 months of 2017 and the related mathematical models.
Table 12. Investigating the relationship between the number of trips made by bicycle in the 12 months of 2017 and the related mathematical models.
ParameterEstimatedtSig.Upper CILower CI
Interceptα+61302.8400.000+6727+5874
Tmeanβ1+9971.91.7580.000+10,251+9746
Hmeanβ2+800.1020.000+142+39
Pmeanβ3−180−0.6470.000−324−64
DLmeanβ4+1358.251.0680.000+1516+1241
WSmeanβ5−15,582−0.2660.000−28,321−12,768
Tmean2γ−317.241−0.1920.000−451−283
Table 13. A monthly average of the predicted number of bicycle trips made in Hamburg in 2017.
Table 13. A monthly average of the predicted number of bicycle trips made in Hamburg in 2017.
YearJanFebMarAprMayJun
Bicycle volume 201717,103,71018,853,28626,872,94530,570,71337,959,18938,521,977
JulAugSepOctNovDec
39,588,69835,168,67029,305,06323,432,66818,436,35914,337,612
Table 14. Correlation between the number of bicycle trips made in 2017 and the number of predicted bicycle trips in 2017, based on Table 13.
Table 14. Correlation between the number of bicycle trips made in 2017 and the number of predicted bicycle trips in 2017, based on Table 13.
CorrelationsBicycle Trips in 2017Predicted Bicycle Trips 2017
Bicycle trips in 2017Pearson Correlation10.971 **
Sig. (2-tailed) 0.000
N1212
Predicted bicycle trips 2017Pearson Correlation0.971 **1
Sig. (2-tailed)0.000
N1212
** The correlation is significant at the 0.01 level (2-tailed).
Table 15. Regression coefficients for monthly bicycle trip volume in 2017 (R2 = 0.942).
Table 15. Regression coefficients for monthly bicycle trip volume in 2017 (R2 = 0.942).
Model Summary
R *R Square **Adjusted R Square ***
0.9710.9420.929
* −1 < R < +1; ** 0 < R2 < 1; *** 0 < Adjusted R2 < 1.
Table 16. The monthly average of the predicted number of bicycle trips made in Hamburg in 2019.
Table 16. The monthly average of the predicted number of bicycle trips made in Hamburg in 2019.
YearJanFebMarAprMayJun
Bicycle trip volume 201915,976,84919,834,91925,413,94532,117,80937,542,24439,943,562
JulAugSepOctNovDec
39,621,24936,079,25629,451,97024,823,44319,225,28715,937,756
Table 17. Correlation between the number of bicycle trips made in 2019 and the number of predicted bicycle trips in 2019.
Table 17. Correlation between the number of bicycle trips made in 2019 and the number of predicted bicycle trips in 2019.
CorrelationsBicycle Trips in 2019Predicted Bicycle Trips 2019
Bicycle trips in 2019Pearson Correlation10.961 **
Sig. (2-tailed) 0.000
N1212
Predicted bicycle trips 2019Pearson Correlation0.961 **1
Sig. (2-tailed)0.000
N1212
** The correlation is significant at the 0.01 level (2-tailed).
Table 18. Regression coefficients for monthly bicycle trip volume in 2019 (R2 = 0.942).
Table 18. Regression coefficients for monthly bicycle trip volume in 2019 (R2 = 0.942).
Model Summary
R *R Square **Adjusted R Square ***Std. Error of the Estimate
0.9610.9240.9072,840,489.903
* −1 < R < +1; ** 0 < R2 < 1; *** 0 < Adjusted R2 < 1.
Table 19. The monthly average of the predicted number of bicycle trips made in Hamburg in 2021 and 2023.
Table 19. The monthly average of the predicted number of bicycle trips made in Hamburg in 2021 and 2023.
Predicted bicycle trips in 2021JanFebMarAprMayJun
14,562,75215,688,31023,867,03925,587,01327,548,57138,396,805
JulAugSepOctNovDec
37,699,41033,396,48137,925,34028,449,63722,638,77413,560,224
Predicted bicycle trips in 2023JanFebMarAprMayJun
17,303,03918,825,73223,041,59128,689,90935,425,89037,954,161
JulAugSepOctNovDec
28,203,36231,957,85336,372,01121,119,85616,638,13611,158,314
Table 20. Monthly correlations between observed and predicted number of cycling trips (2021 and 2023).
Table 20. Monthly correlations between observed and predicted number of cycling trips (2021 and 2023).
CorrelationsBicycle Trips in 2021Predicted Bicycle Trips in 2021
Bicycle trips made in 2021Pearson Correlation10.945 **
Sig. (2-tailed) 0.000
N1212
Predicted bicycle trips in 2021Pearson Correlation0.945 **1
Sig. (2-tailed)0.000
N1212
CorrelationsBicycle trips in 2023Predicted bicycle trips in 2023
Bicycle trips made in 2023Pearson Correlation10.959 **
Sig. (2-tailed) 0.000
N1212
Predicted bicycle trips in 2023Pearson Correlation0.959 **1
Sig. (2-tailed)0.000
N1212
** The correlation is significant at the 0.01 level (2-tailed).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Falah, N.; Falah, N.; Solis-Guzman, J. Weather-Driven Cycling: Developing a Predictive Model for Urban Bicycle Usage Based on Five Key Weather Factors. Urban Sci. 2025, 9, 41. https://doi.org/10.3390/urbansci9020041

AMA Style

Falah N, Falah N, Solis-Guzman J. Weather-Driven Cycling: Developing a Predictive Model for Urban Bicycle Usage Based on Five Key Weather Factors. Urban Science. 2025; 9(2):41. https://doi.org/10.3390/urbansci9020041

Chicago/Turabian Style

Falah, Nahid, Nadia Falah, and Jaime Solis-Guzman. 2025. "Weather-Driven Cycling: Developing a Predictive Model for Urban Bicycle Usage Based on Five Key Weather Factors" Urban Science 9, no. 2: 41. https://doi.org/10.3390/urbansci9020041

APA Style

Falah, N., Falah, N., & Solis-Guzman, J. (2025). Weather-Driven Cycling: Developing a Predictive Model for Urban Bicycle Usage Based on Five Key Weather Factors. Urban Science, 9(2), 41. https://doi.org/10.3390/urbansci9020041

Article Metrics

Back to TopTop