Landing page for Number of crashes resulting in fatalities or serious injuries caused by the top contributing behaviors (speeding, distracted driving, impaired driving, failure to yield) (M.D.2)

This dataset supports measure M.D.2 of SD 2023. The original source of the data is the Texas Department of Transportation supplemented by analysis from the Austin Transportation Department. Each row represents the number of crashes resulting in fatalities or injuries due to the top contributing factors for a year. This dataset can be used to understand the trends in the number and percentages of crashes resulting in serious injuries or fatalities caused by the top contributing factors. View more details and insights related to this measure on the story page : https://data.austintexas.gov/stories/s/9ssh-bavk

United States Life, Accident & Health and Fraternal Entities: Net Investment Yield (Annualized) data was reported at 4.400 % in Dec 2024. This records a decrease from the previous number of 4.800 % for Sep 2024. United States Life, Accident & Health and Fraternal Entities: Net Investment Yield (Annualized) data is updated quarterly, averaging 4.550 % from Mar 2015 (Median) to Dec 2024, with 40 observations. The data reached an all-time high of 5.400 % in Sep 2020 and a record low of 3.900 % in Dec 2022. United States Life, Accident & Health and Fraternal Entities: Net Investment Yield (Annualized) data remains active status in CEIC and is reported by National Association of Insurance Commissioners. The data is categorized under Global Database’s United States – Table US.RG023: Life, Accident & Health and Fraternal Entities.

View more details and insights related to this data set on the story page: https://data.austintexas.gov/stories/s/9ssh-bavk

The 2020 recession did not follow the trend of previous recessions in the United States because only six months elapsed between the yield curve inversion and the 2020 recession. Over the last five decades, 12 months, on average, has elapsed between the initial yield curve inversion and the beginning of a recession in the United States. For instance, the yield curve inverted initially in January 2006, which was 22 months before the start of the 2008 recession. A yield curve inversion refers to the event where short-term Treasury bonds, such as one or three month bonds, have higher yields than longer term bonds, such as three or five year bonds. This is unusual, because long-term investments typically have higher yields than short-term ones in order to reward investors for taking on the extra risk of longer term investments. Monthly updates on the Treasury yield curve can be seen here.

This paper utilizes data on subjective probabilities to study the impact of the stock market crash of 2008 on households' expectations about the returns on the stock market index. We use data from the Health and Retirement Study that was fielded in February 2008 through February 2009. The effect of the crash is identified from the date of the interview, which is shown to be exogenous to previous stock market expectations. We estimate the effect of the crash on the population average of expected returns, the population average of the uncertainty about returns (subjective standard deviation), and the cross-sectional heterogeneity in expected returns (disagreement). We show estimates from simple reduced-form regressions on probability answers as well as from a more structural model that focuses on the parameters of interest and separates survey noise from relevant heterogeneity. We find a temporary increase in the population average of expectations and uncertainty right after the crash. The effect on cross-sectional heterogeneity is more significant and longer lasting, which implies substantial long-term increase in disagreement. The increase in disagreement is larger among the stockholders, the more informed, and those with higher cognitive capacity, and disagreement co-moves with trading volume and volatility in the market.

M.D.2 - Number and percentage of crashes resulting in fatalities or serious injuries caused by the top contributing behaviors (speeding, distracted driving, impaired driving, failure to yield)

A dataset illustrating the most common causes of car accidents along with their corresponding percentage contributions to total accidents. The causes include distracted driving, speeding, driving under the influence, reckless driving, running red lights, driver fatigue, and failure to yield.

Crash Simulations of a Racing Kart Frame Model This dataset contains results for several crash simulations of the frame of a racing kart colliding against a rigid wall. The wall and the frame itself are modeled as finite element models, implemented in the commercial software tool LS-Dyna. The latter comprises 9314 nodes, each possessing 3 translational degrees of freedom. The simulated scenario involves the kart colliding against a rigid wall, with the impact speed varying between 5 and 30 m/s, the impact angle between -45 and 45 degrees and the yield stress between 168 and 758 MPa. The impact angle is the angle between the wall normal and the orientation of the kart, while the yield stress influences the effective plastic stress-strain curve of the kart material. This curve matches that of typical steel, but is adjusted based on the unique yield stress of each simulation. Each crash simulation covers a time span of 0.003 seconds with a sampling interval of 0.3 milliseconds, resulting in 101 samples per simulation. A total of 128 parameter combinations were generated with Halton sequences. Moreover, the source code of the finite element model itself, written for the commercial simulation software LS-DYNA, is included as well. Content Model * input files for the FE simulation software LS-DYNA containing the model description Kart Dataset * simulation results containing the node displacements and simulation parameters. The units are [N,m,s].

This paper utilizes data on subjective probabilities to study the impact of the stock market crash of 2008 on households' expectations about the returns on the stock market index. We use data from the Health and Retirement Study that was fielded in February 2008 through February 2009. The effect of the crash is identified from the date of the interview, which is shown to be exogenous to previous stock market expectations. We estimate the effect of the crash on the population average of expected returns, the population average of the uncertainty about returns (subjective standard deviation), and the cross-sectional heterogeneity in expected returns (disagreement). We show estimates from simple reduced-form regressions on probability answers as well as from a more structural model that focuses on the parameters of interest and separates survey noise from relevant heterogeneity. We find a temporary increase in the population average of expectations and uncertainty right after the crash. The effect on cross-sectional heterogeneity is more significant and longer lasting, which implies substantial long-term increase in disagreement. The increase in disagreement is larger among the stockholders, the more informed, and those with higher cognitive capacity, and disagreement co-moves with trading volume and volatility in the market.

As of December 30, 2024, 14 economies reported a negative value for their ten year minus two year government bond yield spread: Ukraine with a negative spread of 1,370 percent; Turkey, with a negative spread of 1332 percent; Nigeria with -350 percent; and Russia with -273 percent. At this time, almost all long-term debt for major economies was generating positive yields, with only the most stable European countries seeing smaller values. Why is an inverted yield curve important? Often called an inverted yield curve or negative yield curve, a situation where short term debt has a higher yield than long term debt is considered a main indicator of an impending recession. Essentially, this situation reflects an underlying belief among a majority of investors that short term interest rates are about to fall, with the lowering of interest rates being the orthodox fiscal response to a recession. Therefore, investors purchase safe government debt at today's higher interest rate, driving down the yield on long term debt. In the United States, an inverted yield curve for an extended period preceded (almost) all recent recessions. The exception to this is the economic downturn caused by the coronavirus (COVID-19) pandemic – however, the U.S. ten minus two year spread still came very close to negative territory in mid-2019. Bond yields and the coronavirus pandemic The onset of the coronavirus saw stock markets around the world crash in March 2020. This had an effect on bond markets, with the yield of both long term government debt and short term government debt falling dramatically at this time – reaching negative territory in many countries. With stock values collapsing, many investors placed their money in government debt – which guarantees both a regular interest payment and stable underlying value - in contrast to falling share prices. This led to many investors paying an amount for bonds on the market that was higher than the overall return for the duration of the bond (which is what is signified by a negative yield). However, the calculus is that the small loss taken on stable bonds is less that the losses likely to occur on the market. Moreover, if conditions continue to deteriorate, the bonds may be sold on at an even higher price, partly offsetting the losses from the negative yield.

The spread between 10-year and two-year U.S. Treasury bond yields reached a positive value of 0.1 percent in November 2024. The 10-year minus two-year Treasury bond spread is generally considered to be an advance warning of severe weakness in the stock market. Negative spreads occurred prior to the recession of the early 1990s, the tech-bubble crash in 2000-2001, and the financial crisis of 2007-2008.

This analysis presents a rigorous exploration of financial data, incorporating a diverse range of statistical features. By providing a robust foundation, it facilitates advanced research and innovative modeling techniques within the field of finance.

Nifty 50: Climb or Crash?

Financial data:

• Historical daily stock prices (open, high, low, close, volume)

• Fundamental data (e.g., market capitalization, price to earnings P/E ratio, dividend yield, earnings per share EPS, price to earnings growth, debt-to-equity ratio, price-to-book ratio, current ratio, free cash flow, projected earnings growth, return on equity, dividend payout ratio, price to sales ratio, credit rating)

• Technical indicators (e.g., moving averages, RSI, MACD, average directional index, aroon oscillator, stochastic oscillator, on-balance volume, accumulation/distribution A/D line, parabolic SAR indicator, bollinger bands indicators, fibonacci, williams percent range, commodity channel index)

Machine learning features:

• Feature engineering based on financial data and technical indicators

• Sentiment analysis data from social media and news articles

• Macroeconomic data (e.g., GDP, unemployment rate, interest rates, consumer spending, building permits, consumer confidence, inflation, producer price index, money supply, home sales, retail sales, bond yields)

Potential Applications:

• Stock price prediction

• Portfolio optimization

• Algorithmic trading

• Market sentiment analysis

• Risk management

Use Cases:

• Researchers investigating the effectiveness of machine learning in stock market prediction

• Analysts developing quantitative trading Buy/Sell strategies

• Individuals interested in building their own stock market prediction models

• Students learning about machine learning and financial applications

Additional Notes:

• The dataset may include different levels of granularity (e.g., daily, hourly)

• Data cleaning and preprocessing are essential before model training

• Regular updates are recommended to maintain the accuracy and relevance of the data

With fully automated driving systems (ADS; SAE level 4) ride-hailing services expanding in the U.S., we are now approaching an inflection point in the history of vehicle safety assessment. The process of retrospectively evaluating ADS safety impact (as seen with seatbelts, airbags, electronic stability control, etc.) can start to yield statistically credible conclusions. An ADS safety impact measurement requires a comparison to a “benchmark” crash rate. Most benchmarks generated to-date have focused on the current human-driven fleet, which enable researchers to understand the impact of the introduced ADS technology on the current crash record status quo. This study aims to address, update, and extend the existing literature by leveraging police-reported crashes to generate human crash rates for multiple geographic areas with current ADS deployments. Methods: All of the data leveraged is publicly accessible, and the benchmark determination methodology is intended to be repeatable and transparent. Generating a benchmark that is comparable to ADS crash data is associated with certain challenges, including data selection, handling underreporting and reporting thresholds, identifying the population of drivers and vehicles to compare against, choosing an appropriate severity level to assess, and matching crash and mileage exposure data. Consequently, we identify essential steps when generating benchmarks, and present our analyses amongst a backdrop of existing ADS benchmark literature. One analysis presented is the usage of established underreporting correction methodology to publicly available human driver police-reported data to improve comparability to publicly available ADS crash data. We also identified several important crash rate dependencies (geographic region, road type, and vehicle type), and show how failing to account for these features in ADS comparisons can bias results. Working with police-reported crash data to create crash rate benchmarks is fraught with challenges. Researchers should be cautious in their selection of crash rate benchmarks. We present these challenges, discuss their consequences, and provide analytical guidance for addressing them. This body of work aims to contribute to the ability of the community - researchers, regulators, industry, and experts - to reach consensus on how to estimate accurate benchmarks.

Bicycle safety has emerged as a pressing concern within the vulnerable transportation community. Numerous studies have been conducted to identify the significant factors that contribute to the severity of cyclist injuries, yet the findings have been subject to uncertainty due to unobserved heterogeneity and class imbalance. This research aims to address these issues by developing a model to examine the impact of key factors on cyclist injury severity, accounting for data heterogeneity and imbalance. To incorporate unobserved heterogeneity, a total of 3,895 bicycle accidents were categorized into three homogeneous sub-accident clusters using Latent Class Cluster Analysis (LCA). Additionally, five over-sampling techniques were employed to mitigate the effects of data imbalance in each accident cluster category. Subsequently, Bayesian Network (BN) structure learning algorithms were utilized to construct 32 BN models after pairing the accident data from the four accident cluster types before and after sampling. The optimal BN models for each accident cluster type provided insights into the key factors associated with cyclist injury severity. The results indicate that the key factors influencing serious cyclist injuries vary heterogeneously across different accident clusters. Female cyclists, adverse weather conditions such as rain and snow, and off-peak periods were identified as key factors in several subclasses of accident clusters. Conversely, factors such as the week of the accident, characteristics of the trafficway, the season, drivers failing to yield to the right-of-way, distracted cyclists, and years of driving experience were found to be key factors in only one subcluster of accident clusters. Additionally, factors such as the time of the crash, gender of the cyclist, and weather conditions exhibit varying levels of heterogeneity across different accident clusters, and in some cases, exhibit opposing effects.

CSV that contains 1.000 records of interurban road accidents with casualties in Spain between 2016 and 2021. If you are interested in the whole country dataset, please do not hesitate to contact me and I will forward it to you.

Data source of each record is the Spanish General Directorate of Traffic (DGT) and each row describes an accident by the following fields:

• secuencial (int): the unique identifier for an occurrence record in DGT. • anyo (int): the four-digit year of the event date. • mes (int): the month as integer of the event date (encoded). • nombre_mes (str): the month name of the event date (decoded). • dia_semana (int): the integer day of the week when the accident occurred (encoded). • nombre_dia_semana (str): the name of the day when the accident occurred (decoded). • hora (int): the reported hour of the collision in 24-hour notation. • cod_provincia (int): the province code from INE where the accident is registered (encoded). • nombre_provincia (str): the province name where the accident is registered (decoded). • cod_municipio (int): the municipality code from INE where the accident is registered (encoded). • nombre_codigo_municipio (str): the municipality name where the accident is registered (decoded). • isla (str): an integer value to identify a Spanish island if the accident occurred out of the peninsula (encoded). • nombre_isla (str): the island name if applicable to the accident (decoded). • zona (int): an integer value to identify type of road (encoded). • nombre_zona (str): the name of the road type (decoded). • zona_agrupada (int): an integer value to group the road types in urban or interurban (encoded). • nombre_zona_agrupada (str): the group name of the road types (decoded). • carretera (str): the road attending to the national road numbering system in Spain where the accident is located. • km (int): the kilometre point of the road where the accident is located. • sentido_1f (int): the vehicle’s direction of traffic reported as integer when the accident occurred (encoded). • nombre_sentido (str): the vehicle’s direction of traffic reported when the accident occurred (decoded). • titularidad_via (int): the road ownership type as integer (encoded). • nombre_titularidad_via (str): the road ownership type description (decoded). • tipo_via (int): the type of road as integer attending to the project road classification (encoded). • nombre_tipo_via (str): the type of road description attending to the project road classification (decoded). • tipo_accidente (int): an integer value to identify the collision type and traffic context (encoded). • nombre_tipo_accidente (str): the description of the collision type and traffic context (decoded). • total_mu24h (int): the total number of fatalities registered in the accident, computed to 24 hours. • total_hg24h (int): the total number of hospitalised casualties recorded in the accident, counted over 30 days. • total_hl24h (int): the total number of non-hospitalised casualties recorded in the accident, computed to 24 hours. • total_victimas_24h (int): the total number of casualties (fatalities + hospitalised injured + non-hospitalised injured) recorded in the accident, computed over 24 hours. • total_mu30df (int): the total number of fatalities recorded in the accident, computed over 30 days. • total_hg30df (int): the total number of hospitalised casualties recorded in the accident, counted over 30 days. • total_hl30df (int): the total number of non-hospitalised casualties recorded in the accident, counted over 30 days. • total_victimas_30df (int): the total number of casualties (killed + injured in hospital + injured not in hospital) recorded in the accident, counted over 30 days. • total_vehiculos (int): the total number of vehicles involved recorded in the accident. • tot_peat_mu24h (int): the total number of pedestrian fatalities recorded in the accident, computed over 24 hours. • tot_bici_mu24h (int): the total number of cyclists killed recorded in the accident, computed to 24 hours. • tot_ciclo_mu24h (int): the total number of scooter riders killed recorded in the accident, computed on a 24-hour basis. • tot_moto_mu24h (int): the total number of motorcyclist fatalities recorded in the accident, computed on a 24-hour basis. • tot_tur_mu24h (int): the total number of car drivers and passengers killed recorded in the accident, counted to 24 hours. • tot_furg_mu24h (int): the total number of van drivers and passengers killed recorded in the accident, counted over 24 hours. • tot_cam_menos3500_mu24h (int): the total number of drivers and passengers of trucks ≤ 3,500 kg killed in the accident, counted over 24 hours. • tot_cam_mas3500_mu24h (int): the total number of drivers and passengers of trucks > 3,500 kg killed in the accident, counted over 24 hours. • tot_bus_mu24h (int): the total number of bus drivers and passengers fatalities recorded in the accident, counted over 24 hours. • tot_otro_mu24h (int): the total number of drivers and passengers of vehicles not classified in the above types killed in the accident, counted to 24 hours. • tot_sinespecif_mu24h (int): the total number of drivers and passengers of vehicles of unspecified type killed in the accident, counted over 24 hours. • tot_peat_mu30df (int): the total number of pedestrian fatalities recorded in the accident, computed to 30 days. • tot_bici_mu30df (int): the total number of cyclists killed recorded in the accident, counted over 30 days. • tot_ciclo_mu30df (int): the total number of scooted riders killed recorded in the accident, counted over 30 days. • tot_moto_mu30df (int): the total number of motorcyclist fatalities recorded in the accident, counted over 30 days. • tot_tur_mu30df (int): the total number of drivers and passengers of passenger cars killed in the accident, counted over 30 days. • tot_furg_mu30df (int): the total number of van drivers and passengers killed recorded in the accident, counted over 30 days. • tot_cam_menos3500_mu30df (int): the total number of drivers and passengers of trucks ≤ 3,500 kg killed in the accident, counted over 30 days. • tot_cam_mas3500_mu30df (int): the total number of drivers and passengers of trucks > 3,500 kg killed in the accident, counted over 30 days. • tot_bus_mu30df (int): the total number of bus drivers and passengers fatalities recorded in the accident, counted over 30 days. • tot_otro_mu30df (int): the total number of drivers and passengers of vehicles of types not classified in the above killed in the accident, counted over 30 days. • tot_sinespecif_mu30df (int): the total number of drivers and passengers of unspecified type vehicles killed in the crash, counted over 30 days. • nudo (int): an integer to identify whether the collisions occurred in an road junction or not (encoded). • nombre_nudo (str): the description to identify whether the collisions occurred in an road junction or not (decoded). • nudo_info (int): an integer that represents the type of road junction in which the collision occurred (encoded). • nombre_nudo_info (str): the description of the type of road junction in which the collision occurred (decoded). • carretera_cruce (str): the road attending to the national road numbering system in Spain where the accident is located. • priori_norma (int): an integer to identify if the road junction priority is determined by generic traffic rule (encoded). • nombre_priori_norma (str): the text to identify if the road junction priority is determined by generic traffic rule (decoded). • priori_agente (int): an integer to identify if the road junction priority is determined by agent (encoded). • nombre_priori_agente (str): the text to identify if the road junction priority is determined by agent (decoded). • priori_semaforo (int): an integer to identify if the road junction priority is determined by traffic light (encoded). • nombre_priori_semaforo (str): the text to identify if the road junction priority is determined by traffic light (decoded). • priori_vert_stop (int): an integer to identify if the road junction priority is determined by vertical stop sign (encoded). • nombre_priori_vert_stop (str): the text to identify if the road junction priority is determined by vertical stop sign (decoded). • priori_vert_ceda (int): an integer to identify if the road junction priority is determined by vertical yield sign (encoded). • nombre_priori_priori_vert_ceda (str): the text to identify if the road junction priority is determined by vertical yield sign (decoded). • priori_horiz_stop (int): an integer to identify if the road junction priority is determined by horizontal stop signal (encoded). • nombre_priori_horiz_stop (str): the text to identify if the road junction priority is determined by horizontal stop signal (decoded). • priori_horiz_ceda (int): an integer to identify if the road junction priority is determined by horizontal yield sign (encoded). • nombre_priori_horiz_ceda (str): the text to identify if the road junction priority is determined by horizontal yield sign (decoded). • priori_marcas (int): an integer to identify if the road junction priority is determined by road markings (encoded). • nombre_priori_marcas (str): the text to identify if the road junction priority is determined by road markings

The paper questions whether the currency crash risk functions as a pricing factor in carry trade activies. We explore the currency crash risk measures can provide insight into understanding the dynamics of carry trade strategies and their associated returns. Our analysis using different currency crash risk measures documents that the carry returns are, indeed, driven by the crash risk of investment currency. By comparing the returns of unhedged and crash-hedged carry trade strategies, we assess the return that investors seek in response to currency crash risk. Our findings demonstrate that, on average, 62 percent of the carry returns can be attributed to the currency crash risk

Automotive Crash Test Dummy Market Outlook

The global automotive crash test dummy market size in 2023 was estimated to be USD 450 million and is projected to reach USD 710 million by 2032, growing at a CAGR of 5.2% during the forecast period. The increasing emphasis on vehicle safety, driven by stringent government regulations and consumer demand, is a significant growth factor for this market. Technological advancements in automotive design and an increase in vehicle production worldwide are also contributing to this robust growth.

One of the primary growth factors for the automotive crash test dummy market is the stringent safety regulations imposed by various governments. These regulations mandate comprehensive testing of vehicles to ensure they meet safety standards, thus driving the demand for advanced crash test dummies. The rising awareness among consumers about vehicle safety is another contributing factor. Consumers are now more inclined to purchase vehicles that promise higher safety ratings, prompting manufacturers to invest in better crash test technologies.

Additionally, technological advancements in the design and functionality of crash test dummies have been pivotal in market growth. Modern crash test dummies are now equipped with sophisticated sensors and data acquisition systems that provide detailed insights into the impacts and stresses experienced during a crash. This has enhanced the accuracy and reliability of crash simulations, making them indispensable tools for automotive manufacturers. The ongoing research and development activities in this sector are expected to yield even more advanced and specialized dummies, further propelling market growth.

The increase in global vehicle production is another significant driver for the automotive crash test dummy market. Emerging economies, particularly in the Asia Pacific region, have seen a surge in automobile manufacturing due to urbanization, rising disposable incomes, and favorable government policies. This has led to an increased demand for crash test dummies as automotive manufacturers strive to meet both regulatory requirements and consumer expectations for safer vehicles.

Regionally, North America and Europe have been leading the market due to their stringent safety regulations and the presence of major automobile manufacturers. However, the Asia Pacific region is expected to witness the highest growth rate during the forecast period. This can be attributed to the rapid industrialization, increasing vehicle production, and supportive government initiatives aimed at enhancing vehicle safety standards. Countries such as China, Japan, and India are anticipated to be major contributors to this growth.

Type Analysis

The automotive crash test dummy market is segmented by type into adult, child, and infant categories. Adult dummies are the most widely used due to the higher proportion of adult vehicle occupants. These dummies are designed to represent a typical adult with standard male and female versions, each equipped with sensors to measure different types of impact. The detailed data they provide helps manufacturers design vehicles that offer better protection to adult occupants during crashes.

Child crash test dummies play a crucial role in enhancing the safety of younger passengers. These dummies are specifically designed to simulate the body structure and weight distribution of children of various age groups. With increasing awareness about child safety in vehicles, the demand for child crash test dummies is on the rise. Regulatory bodies require comprehensive testing of child restraint systems, which necessitates the use of these specialized dummies.

Infant crash test dummies are another critical segment, reflecting the need to ensure the safety of the youngest passengers. These dummies are designed to mimic the physical characteristics of infants, including weight, length, and body mechanics. They are primarily used to test infant car seats and other restraint systems. The growing focus on providing adequate protection for infants in vehicles is driving the demand for these dummies.

Each type of crash test dummy undergoes rigorous development and testing to ensure it accurately represents the demographics it is intended for. The data collected from these tests are crucial for improving vehicle designs and s

The top locations where reported collisions occurred between bicyclists and motor vehicles have been identified. The crash cluster analysis methodology for the top bicyclist clusters uses a fixed meter search distance of 100 meters (328 ft.) to merge crash clusters together. Located crashes between motor vehicles and bicyclists were identified by using the non-motorist type code within the CDS database (which may yield different results from using most harmful event, first harmful event, or sequence of events data fields). Furthermore, the methodology uses the Equivalent Property Damage Only (EPDO) weighting to rank the clusters. However, because of the relatively small number of reported bicyclists crashes in the crash data file, the clustering analysis used crashes from the ten year period from 2012-2021. Additionally, due to the larger geographic area encompassed by the bicyclist crash clusters, it was difficult to name them so they were left unnamed but can be viewed spatially.

Bicycle safety has emerged as a pressing concern within the vulnerable transportation community. Numerous studies have been conducted to identify the significant factors that contribute to the severity of cyclist injuries, yet the findings have been subject to uncertainty due to unobserved heterogeneity and class imbalance. This research aims to address these issues by developing a model to examine the impact of key factors on cyclist injury severity, accounting for data heterogeneity and imbalance. To incorporate unobserved heterogeneity, a total of 3,895 bicycle accidents were categorized into three homogeneous sub-accident clusters using Latent Class Cluster Analysis (LCA). Additionally, five over-sampling techniques were employed to mitigate the effects of data imbalance in each accident cluster category. Subsequently, Bayesian Network (BN) structure learning algorithms were utilized to construct 32 BN models after pairing the accident data from the four accident cluster types before and after sampling. The optimal BN models for each accident cluster type provided insights into the key factors associated with cyclist injury severity. The results indicate that the key factors influencing serious cyclist injuries vary heterogeneously across different accident clusters. Female cyclists, adverse weather conditions such as rain and snow, and off-peak periods were identified as key factors in several subclasses of accident clusters. Conversely, factors such as the week of the accident, characteristics of the trafficway, the season, drivers failing to yield to the right-of-way, distracted cyclists, and years of driving experience were found to be key factors in only one subcluster of accident clusters. Additionally, factors such as the time of the crash, gender of the cyclist, and weather conditions exhibit varying levels of heterogeneity across different accident clusters, and in some cases, exhibit opposing effects.