Crash Patterns Research Articles

The safety impacts of paved shoulder width in Indian four-lane rural highways

The shoulder width, as a geometric element, plays a crucial role in enhancing highway safety. Research from high-income countries indicates that improving shoulders on highways leads to substantial safety benefits. However, the safety effectiveness of paved shoulders for low- and middle-income countries (LMICs) highway contexts has limited evidence. This study evaluated the safety effectiveness of the paved shoulder width on 61 km, four-lane, divided rural intercity highways in India. The first objective was to evaluate highway crash patterns using data from 2016 to 2019. The second objective was to evaluate the safety effectiveness of paved shoulder width using the case-control approach. The findings of this study demonstrate a consistent decline in the likelihood of crashes as the shoulder’s width increases within the range of zero to 2.5 m for the 100 m segment length and zero to 1.7 m for the 500 m segment length. Nevertheless, model estimates indicate an increased crash risk for shoulders wider than 2.5 m. The results also suggested that the odds ratio for paved shoulder widths ranging from no shoulder to 2.5 m is likely to follow the crash modification factor from the highway safety manual. The findings of this study hold significant implications for the design policy of shoulder width on rural highways in LMICs.

Investigating the Interrelationships among Factors Associated with Automated Vehicle Crashes Using Additive Bayesian Network

Although automated vehicles (AVs) were considered a promising solution to enhance traffic safety by eliminating human errors, AV crashes still happen in mixed traffic consisting of human-driven vehicles and AVs. Thus, to reduce AV-involved crashes, it is necessary to understand the factors leading to AV crashes. However, traditional regression-based methods may not reveal a structured relationship among leading factors of AV crashes, which hinders the exploration of countermeasures to AV crashes. Based on the 246 AV crash records collected by the National Highway Traffic Safety Administration, this study investigated the factors associated with AV crashes. An additive Bayesian network (ABN) approach was utilized to construct the topological relationship among potential influential factors of AV crashes, followed by post-ABN regression analyses. Results show that, though AV technologies have developed rapidly in the past few years, rear-end crashes are still dominant among AV-involved crashes, potentially because of the discrepancy in the driving behaviors between AV and human-driven vehicles. The crash type of AV-involved crashes is more related to the pre-crash movements of crash partners than it is to the pre-crash movements of AVs, while crash outcomes (e.g., injury severity) are associated with the environmental factors (e.g., operating entities) and crash-procedure-related factors (e.g., crash type). Findings from this study aid in understanding AV crash patterns, which can inform targeted interventions and technology advancements to improve safety outcomes for all road users.

Analyzing urban traffic crash patterns through spatio-temporal data: A city-level study using a sparse non-negative matrix factorization model with spatial constraints approach

Urban traffic crashes represent a significant challenge affecting public safety and urban mobility worldwide. This study introduces a novel application of Sparse Non-negative Matrix Factorization with spatial constraints to analyze spatio-temporal patterns of traffic crashes at a city level. Using comprehensive crash data from Denver and Manhattan during 2020, we developed and validated a model capable of capturing distinct temporal dynamics and spatial distributions of traffic crashes. Unlike traditional methods, our approach integrates sparsity and spatial constraints, enhancing the model's ability to handle the inherent sparsity and geographical dependencies found in urban traffic data. The results demonstrate the model's effectiveness in identifying high-risk areas and times, providing actionable insights that can inform urban planning and targeted safety interventions. The study underscores the potential of advanced data-driven techniques in urban traffic analysis and contributes to the broader efforts of improving traffic safety through informed decision-making and policy development.

Road crash dynamics in Malaysia: Analysis of trends and patterns

Road crashes represent a significant public health and safety concern globally, and Malaysia is no exception. Understanding the trends and patterns of road crashes is essential for devising effective strategies to mitigate risks and enhance road safety. This study presents a comprehensive analysis of road crash dynamics, focusing on road users, severity patterns and geographical patterns in Malaysia from 2012 to 2022. Data sourced from the Royal Malaysian Police (RMP) are utilized to examine various aspects of road crashes. Road crash trend, geographical patterns, linear trend analysis and K-means clustering are employed to explore patterns of road crash in Malaysia. The findings reveal that motorcycles consistently emerged as the most involved road user. Geographical patterns discovered that Selangor exhibits higher crash number. Linear trend analysis revealed significant upward trends in crash frequency prior to the pandemic, while the number of fatalities resulting from road crash showed a downward trend over the observed period. K-means clustering identified that Selangor recorded high total crashes and high total of fatalities. This study also considers the influence of the Covid-19 pandemic on road crash dynamics, highlighting changes in travel patterns and behaviour. There also have been notable successes, such as the reduction in total fatalities and the effectiveness of targeted interventions via the accomplishments of initiatives of Malaysian Road Safety Plan 2014–2022.

Assessing network-based traffic crash risk using prospective space-time scan statistic method

As car ownership and urbanization continue to rise worldwide, traffic crashes have become growing concerns globally. Measuring crash risk provides insight into understanding crash patterns, which can eventually support proactive transport planning and improve road safety. However, traditional spatial analysis methods for crash risk assessment, such as the hotspot detection method, are mainly focused on identifying areas with higher crash frequency. These methods are subject to critical issues in risk analysis due to ignoring crash impacts and background traffic volume information. Aside from the two issues, current crash risk assessment methods, especially those aiming for cluster detection, are subject to the modified temporal unit problem, referring to the temporal effects (i.e., aggregation, segmentation, and boundary) in cluster detection. To alleviate these issues, this paper applies an emerging hot spot detection method, called the prospective space-time scan statistic (STSS) method, for assessing the crash risk at a refined network scale and over multiple years in a case study of Hartford, Connecticut. By identifying the spatial and temporal clusters of the crash risk, the study can provide evidence for tailoring road safety management strategies in neighborhoods characterized by high crash risk.

An In-Depth Analysis of Road Fatal Crash Patterns and Discussions in Ho Chi Minh City

An In-Depth Analysis of Road Fatal Crash Patterns and Discussions in Ho Chi Minh City

Generation of nighttime pedestrian fatal precrash scenarios at junctions in Tamil Nadu, India, using cluster correspondence analysis

Objective Modern transportation amenities and lifestyles have changed people’s behavioral patterns while using the road, specifically at nighttime. Pedestrian and driver maneuver behaviors change based on their exposure to the environment. Pedestrians are more vulnerable to fatal injuries at junctions due to increased conflict points with vehicles. Generation of precrash scenarios allows drivers and pedestrians to understand errors on the road during driver maneuvering and pedestrian walking/crossing. This study aims to generate precrash scenarios using comprehensive nighttime fatal pedestrian crashes at junctions in Tamil Nadu, India. Methods Though numerous studies were available on identifying pedestrian crash patterns, only some focused on identifying crash patterns at junctions at night. We used cluster correspondence analysis (CCA) to address this research gap to identify the patterns in nighttime pedestrian fatal crashes at junctions. Further, high-risk precrash scenarios were generated based on the positive residual means available in each cluster. This study used crash data from the Road Accident Database Management System of Tamil Nadu State in India from 2009 to 2018. Characteristics of pedestrians, drivers, vehicles, crashes, light, and roads were input to the CCA to find optimal clusters using the average silhouette width, Calinski-Harabasz measure, and objective values. Results CCA found 4 clusters with 2 dimensions as optimal clusters, with an objective value of 3.3618 and a valence criteria ratio of 80.03%. Results from the analysis distinctly clustered the pedestrian precrash behaviors: Clusters 1 and 2 on pedestrian walking behaviors and clusters 3 and 4 on crossing behaviors. Moreover, a hidden pattern was observed in cluster 4, such as transgender drivers involved in fatal pedestrian crashes at junctions at night. Conclusion The generated precrash scenarios may be used to train drivers (novice and inexperienced for nighttime driving), test scenario creation for developing advanced driver/rider assistance systems, hypothesis creation for researchers, and planning of effective strategic interventions for engineers and policymakers to change pedestrian and driver behaviors toward sustainable safety on Indian roads.

Adjusting vehicle secondary safety ratings to account for crash-avoidance technology fitment using real-world crash and injury data

Objectives Publicizing safety ratings of vehicles can motivate manufacturers to prioritize safety and help consumers choose safer vehicles, leading to safer fleets. The benefits of primary safety technologies that prevent crash occurrence are not currently incorporated in current ratings in a way that values their safety benefits consistently. We aimed to propose a method for assigning weights for each safety technology to account for established safety benefits using published effectiveness and prevalence from real-life data. Methods To illustrate this method, we present a worked example calculated using crash and injury data from Australia and New Zealand. The method proposed attenuates the weights for given safety technologies where two or more safety technologies fitted to the same vehicle are effective for the same types of crashes. Results In the worked example using Australasian data, large SUVs were estimated to have the largest safety increment from the fitment of all the technologies considered compared to vehicles without these primary safety technologies, with an almost 17% reduction in crash occurrence. Cars with all the technologies fitted had estimated average crash reduction of between 11% and 12%. Conclusions Different market groups have different crash patterns, so the safety attributable to safety technology fitment differs at the market group level. This study presents an approach for providing a summary measure of crash avoidance according to the fitment of safety technologies. If this measure is combined with an estimate of secondary safety (whether derived from existing crash and injury data or from new car crash assessment programs), the combined estimate then represents the important elements of safety provided by the vehicle. The methods presented here form a rational basis for assigning safety ratings to represent the benefits of swiftly developing safety technologies.

A spatial autocorrelation analysis of road traffic crash by severity using Moran’s I spatial statistics: A comparative study of Addis Ababa and Berlin cities

Methodological advancements in road safety research reveal an increasing inclination toward integrating spatial approaches in hot spot identification, spatial pattern analysis, and developing spatially lagged models. Previous studies on hot spot identification and spatial pattern analysis have overlooked crash severities and the spatial autocorrelation of crashes by severity, missing valuable insights into crash patterns and underlying factors. This study investigates the spatial autocorrelation of crash severity by taking two capital cities, Addis Ababa and Berlin, as a case study and compares patterns in low and high-income countries. The study used three-year crash data from each city. It employed the average nearest neighbor distance (ANND) method to determine the significance of spatial clustering of crash data by severity, Global Moran's I to examine the statistical significance of spatial autocorrelation, and Local Moran's I to identify significant cluster locations with High-High (HH) and Low-Low (LL) crash severity values. The ANND analysis reveals a significant clustering of crashes by severity in both cities, except in Berlin's fatal crashes. However, different Global Moran's I results were obtained for the two cities, with a strong and statistically significant value for Addis Ababa compared to Berlin. The Local Moran's I result indicates that the central business district and residential areas have LL values, while the city's outskirts exhibit HH values in Addis Ababa. With some persistent HH value locations, Berlin's HH and LL grid clusters are intermingled on the city's periphery. Socio-economic factors, road user behavior and roadway factors contribute to the difference in the result. Nevertheless, it is interesting to note the similarity of significant HH value locations on the outskirts of both cities. Finally, the results are consistent with previous studies and indicate the need for further investigation in other locations.

Fatal Crash Occurrence Prediction and Pattern Evaluation by Applying Machine Learning Techniques

Background Highway safety remains a significant issue, with road crashes being a leading cause of fatalities and injuries. While several studies have been conducted on crash severity, few have analyzed and predicted specific types of crashes, such as fatal crashes. Identifying the key factors associated with fatal crashes and predicting their occurrence can help develop effective preventative measures. Objective This study intended to develop cluster analysis and ML-based models using crash data to extract the prominent factors behind fatal crash occurrences and analyze the inherent pattern of variables contributing to fatal crashes. Methods Several branches and categories of supervised ML models have been implemented for fatality prediction and their results have been compared. SHAP analysis was conducted using the ML model to explore the contributing factors of fatal crashes. Additionally, the underlying hidden patterns of fatal crashes have been evaluated using K-means clustering, and specific fatal crash scenarios have been extracted. Results The deep neural networks model achieved 85% accuracy in predicting fatal crashes in Kansas. Factors, such as speed limits, nighttime, darker road conditions, two-lane highways, highway interchange areas, motorcycle and tractor-trailer involvement, and head-on collisions were found to be influential. Moreover, the clusters were able to discern certain scenarios of fatal crashes. Conclusion The study can provide a clear image of the important factors related to fatal crashes, which can be utilized to create new safety protocols and countermeasures to reduce fatal crashes. The results from cluster analysis can facilitate transportation professionals with representative scenarios, which will benefit in identifying potential fatal crash conditions.

An Econometric Analysis to Explore the Temporal Variability of the Factors Affecting Crash Severity Due to COVID-19

This study utilizes multilevel ordinal logistic regression (M-OLR), an approach that accounts for spatial heterogeneity, to assess the dynamics of crash severity in Virginia, USA, over the years 2018 to 2023. This period was notably influenced by the COVID-19 pandemic and its associated stay-at-home orders, which significantly altered traffic behaviors and crash severity patterns. This study aims to evaluate the pandemic’s impact on crash severity and examine the consequent changes in driver behaviors. Despite a reduction in total crashes, a worrying increase in the proportion of severe injuries is observed, suggesting that less congested roads during the pandemic led to riskier driving behaviors, notably increased speed violations. This research also highlights heightened risks for vulnerable road users such as pedestrians, cyclists, and motorcyclists, with changes in transportation habits during the pandemic leading to more severe crashes involving these groups. Additionally, this study emphasizes the consistent influence of environmental and roadway features, like weather conditions and traffic signals, in determining crash outcomes. These findings offer vital insights for road safety policymakers and urban planners, indicating the necessity of adaptive road safety strategies in response to changing societal norms and behaviors. The research underscores the critical role of individual behaviors and mental states in traffic safety management and advocates for holistic approaches to ensure road safety in a rapidly evolving post-pandemic landscape.

Severity Analysis of Secondary Crashes on High-Speed Roadways: Pattern Recognition Using Association Rule Mining

Secondary crashes (SCs) are a major concern, posing additional safety threats to both non-involved vehicles and incident responders. The objective of this study was to identify the affiliated factors contributing to SCs on roadways with a speed limit of 55 mph or above. Traditional police-investigated crash dataset was analyzed, spanning more than four years (January 2016–February 2020) for the entire state of Alabama. As the crash database did not directly include information on SCs and did not allow for linking a primary crash with a subsequent SC, a data extraction process was developed to identify SCs and understand their characteristics. Association rule mining (ARM) was applied to identify crash patterns based on maximum injury severity levels. The generated rules were filtered based on support, confidence, and lift, and then validated by the lift increase criterion. The results revealed complex relationships between risk factors and severity of SCs. In relation to SCs with injuries, single-vehicle crashes were frequently observed during peak hours and when drivers swerved to avoid objects/persons/vehicles. In contrast, concerning SCs with possible/no injuries, single-vehicle collisions were more likely to occur when drivers failed to notice objects/persons/vehicles and were involved in speeding. On urban interstates, single-vehicle SCs were frequently associated with injuries, while rear-end SCs were often linked to possible/no injuries. The findings of this study can be helpful in enhancing existing traffic incident management programs to mitigate the occurrence of SCs.

Crash Contributing Factors and Patterns Associated with Fatal Truck-Involved Crashes in Bangladesh: Findings from the Text Mining Approach

Despite extensive research on traffic injury severities, relatively little is known about the factors contributing to truck-involved crashes in developing countries, especially in the context of Bangladesh. Because of the unavailability of authentic crash data sources, this study collected data from alternative sources such as online English news media reports. The current study prepared a database of 144 truck-involved fatal crash reports during the period of 12 months (January 2021 to December 2021). The crash reports contain a bag of 15,300 words. Several state-of-the-art text mining tools were utilized to identify crash patterns, including word cloud analysis, word frequency analysis, word co-occurrence network analysis, rapid automatic keyword extraction, and topic modeling. The analysis revealed several important crash contributing factors, such as the type of vehicle involved (auto-rickshaw, bus, van, motorcycle), the manner of collision (head-on), the time of the day (morning, night), driver behavior (speeding, overtaking, wrong-way driving), and environmental factors (dense fog). In addition, “coming from opposite direction” and “head-on collision” are two important sequences of events in truck-involved crashes. Truck drivers are also involved in crashes with trains at rail crossings. The findings of this research can assist policymakers in identifying crash avoidance strategies to lower truck-related crashes in Bangladesh.

Characteristics of Rear-End Collisions: A Comparison between Automated Driving System-Involved Crashes and Advanced Driving Assistance System-Involved Crashes

With the increasing number of vehicles equipped with automated driving systems (ADS) being tested on public roads and the expanding market share of vehicles equipped with advanced driving assistance systems (ADAS), the number of ADS- or ADAS-involved crashes increases. Thus, it is necessary to investigate the distribution of ADS- and ADAS-involved crashes and the factors leading to them. The rear-end collision has been found to dominate ADS-involved crashes. However, no research has explored the conditions when ADS-involved rear-end collisions are more likely to happen and no research has investigated ADAS-involved rear-end crashes. Based on 130 ADS-involved crashes and 84 ADAS-involved crashes extracted from a dataset collected by the National Highway Traffic Safety Administration (NHTSA) between July 2021 and May 2022, this study explored the crash patterns, especially rear-end crashes, of ADS- and ADAS-controlled vehicles. Results show that rear-end collisions dominate both ADS- and ADAS-involved crashes, especially ADAS-involved crashes. The type of both ADS-involved and ADAS-involved crashes was affected by the speed of the ego-vehicle relative to the posted speed limit. Further, the type of ADS-involved crash was affected by the pre-crash movement of the crash partner, while the type of ADAS-involved crash was further associated with the road type. The findings can provide insights into the design of ADAS and ADS control algorithms, the external human-machine interface design of the vehicles with ADS or ADAS, and the training program of human road users to improve traffic safety in mixed traffic.

A comparison of patterns and contributing factors of ADAS and ADS involved crashes

Crashes involving Advanced Driver Assistance Systems (ADAS) and Automated Driving Systems (ADS) have been increasing in recent years. Understanding the characteristics of these crashes can guide the optimization of driving automation systems and the policies improving the safety of mixed traffic. However, due to the limited available data, the crashes of ADS- and ADAS-controlled vehicles are still under-investigated. Thus, utilizing the latest National Highway Traffic Safety Administration crash reports, our study explores the patterns and contributing factors of ADAS- and ADS-involved crashes. The sequences of events leading to crashes were extracted from the reports and then categorized into five clusters. Next, for incomplete records, a non-parametric imputation method was applied based on Random Forest. Finally, logistic regression models were built to explore the factors associated with the crashes. The results show that the automation level, speed limit, and vehicle speed are predictors of crash patterns. At the same time, the crash pattern, combined with incident time, roadway type, roadway surface, and vehicle model year are associated with crash outcomes (i.e. contact area and injury severity). The results indicate that further improvement of the ADS/ADAS control algorithms and driver education may be needed to improve the safety of mixed traffic.

Crashes involving distracted pedestrians: Identifying risk factors and their relationships to pedestrian severity levels and distraction modes

The concept of distracted pedestrians and its impact on highway safety has gained increasing attention in recent years. However, studies focusing exclusively on distracted pedestrian crashes are less pervasive than distracted driving. In addition, most prior studies investigate the harmful effect of cellphone usage while walking, without considering other forms of pedestrian distraction. Also, the existing literature provides limited knowledge on comprehending the affinities between pedestrian distraction and safety consequences. This study aims to reveal the chain of contributing factors involved in distracted pedestrian crashes, considering both pedestrian severity levels and specific distraction-related tasks. Ten years (2010–2019) of related crashes were extracted from the Louisiana Department of Transportation and Development (LADOTD) database, and association rule mining (ARM) was applied to identify the meaningful crash patterns. Different distracting activities of pedestrians were introduced from the narratives of police-investigated crash reports. The study findings exhibit the complex nature of distracted pedestrian crashes by highlighting the intricate relationships between risk factors. On road segments, distracted male pedestrians aged 41–64 were more likely to be fatal/severely injured in dark-not-lighted conditions. Crashes involving pedestrians using electronic devices were often found at intersections. Distractions caused by pets, persons, or objects were strongly associated with crossing segments in rural settings. In-person conversation while standing on roadways in urban residential locations without traffic controls was found to increase vulnerability. Working on vehicles while wearing dark clothes and in dark-not-lighted conditions was identified as an influential factor in crash occurrence. Moreover, careless or inattentive actions of pedestrians while playing on the road segments were associated with a high likelihood of crashes. These study outcomes are crucial in uncovering the coexisting crash characteristics related to distracted pedestrians, which can be helpful in targeting and developing effective educational, design, and enforcement strategies to improve pedestrian safety.

Improving model robustness of traffic crash risk evaluation via adversarial mix-up under traffic flow fundamental diagram

Recent state-of-art crash risk evaluation studies have exploited deep learning (DL) techniques to improve performance in identifying high-risk traffic operation statuses. However, it is doubtful if such DL-based models would remain robust to real-world traffic dynamics (e.g., random traffic fluctuations.) as DL models are sensitive to input changes, where small perturbations could lead to wrong predictions. This study raises the critical robustness issue for crash risk evaluation models and investigates countermeasures to enhance it. By mixing up crash and non-crash samples under the traffic flow fundamental diagram, traffic flow adversarial examples (TF-AEs) were generated to simulate real-world traffic fluctuations. With the developed TF-AEs, model accuracy decreased by 8% and sensitivity dropped by 18%, indicating weak robustness of the baseline model (a convolutional neural network, CNN-based crash risk evaluation model). Then, a coverage-oriented adversarial training method was proposed to improve model robustness in highly imbalanced crash and non-crash situations and various crash risk transition patterns. Experiments showed that the proposed method was effective to improve model robustness as it could prevent 76.5% accuracy drops and 98.9% sensitivity drops against TF-AEs. Finally, the evaluation model outputs’ stability and limitations of the current study are discussed.

Exploring the impact of trip patterns on spatially aggregated crashes using floating vehicle trajectory data and graph Convolutional Networks

In recent years, increased attention has been given to understanding the spatial pattern of crashes in urban areas. Accurately capturing the spatial relationship between crash counts and variables requires extracting hidden information from multiple data sources. In this study, we propose a machine learning model to explore the spatial impact of activity patterns on spatially aggregated crash counts. Our paper introduces a two-step framework: (a) the Latent Dirichlet Allocation (LDA) model, an unsupervised method for mining hidden activity patterns from floating vehicle trajectory data, and (b) the Graph Convolutional Network (GCN) model, which builds the spatial relationship between multi-source data. The data and hidden activity patterns were aggregated into 175 Traffic Analysis Zones (TAZs) in San Francisco using spatial partitioning. The GCN model provided higher prediction accuracy than commonly used machine learning algorithms that did not consider combined spatial relationships and those that only considered traditional vehicle counts data. Furthermore, we used attribution algorithms to obtain the respective weight scores of each factor. Our results reveal that daily vehicle kilometers traveled, road density, population density, commercial activity during weekends, and residential activity during morning peak hours on weekdays are factors associated with crashes.

Investigating the spatial heterogeneity of factors influencing speeding-related crash severities using correlated random parameter order models with heterogeneity-in-means

ABSTRACT Speeding has been acknowledged as a critical determinant in increasing the risk of crashes and their resulting injury severities. This paper employs Global Moran’s I coefficient and local Getis – Ord G* indexes to systematically account for the spatial distribution feature of speeding-related crashes, study the global spatial pattern of speeding-related crashes, and identify severe crash cluster districts. The findings demonstrate that severe speeding-related crashes within the state of Pennsylvania have a spatial clustering trend, where four crash datasets are extracted from four hotspot districts. Two log-likelihood ratio (LR) tests were conducted to determine whether speeding-related crashes classified by hotspot districts should be modeled separately. The results suggest that separate modeling is necessary. To capture the unobserved heterogeneity, four correlated random parameter order models with heterogeneity in means are employed to explore the factors contributing to crash severity involving at least one vehicle speeding.

Real-time crash risk forecasting using Artificial-Intelligence based video analytics: A unified framework of generalised extreme value theory and autoregressive integrated moving average model

With the recent advancements in computer vision and artificial intelligence, traffic conflicts occurring at an intersection and associated traffic characteristics can be obtained at the granular level of a signal cycle in real-time. This capability enables the estimation of the real-time crash risk using sophisticated modelling techniques, e.g., extreme value theory. However, these models are inherently incapable of forecasting the crash risk of future time periods based on the temporal dependency of crash risks. This study proposes a unified framework of extreme value theory and autoregressive integrated moving average models for forecasting crash risks at signalised intersections. At the first level of this framework, a non-stationary generalised extreme value model has been developed to estimate the real-time rear-end crash risk at the signal cycle level using the video data collected from three signalised intersections in Queensland, Australia. To capture the time-varying effect of different traffic conditions on conflict extremes, traffic flow, speed, shockwave area, and platoon ratio covariates are incorporated into the generalised extreme value model. The signal cycle-level crash risks obtained from the first level form a univariate time series, which is modelled using two variants of autoregressive integrated moving average model to forecast the crash risk of future signal cycles. Results reveal that the autoregressive integrated moving average model with exogenous variables outperforms the model without exogenous variables and can forecast the crash risk for the next 30–35 min with reasonable accuracy. Similarly, results also demonstrate that different crash risk patterns within a typical day are accurately predicted. The proposed framework helps identify the spatiotemporal windows where safety gradually deteriorates over time, thus enabling proactive safety assessment.