Bayesian Poisson-lognormal Model Research Articles

Age- and Sex-Specific Association Between Vegetation Cover and Mental Health Disorders: Bayesian Spatial Study.

BackgroundDespite growing evidence that reduced vegetation cover could be a putative risk factor for mental health disorders, the age- and the sex-specific association between vegetation and mental health disorder cases in urban areas is poorly understood. However, with rapid urbanization across the globe, there is an urgent need to study this association and understand the potential impact of vegetation loss on the mental well-being of urban residents.ObjectiveThis study aims to analyze the spatial association between vegetation cover and the age- and sex-stratified mental health disorder cases in the neighborhoods of Toronto, Canada.MethodsWe used remote sensing to detect urban vegetation and Bayesian spatial hierarchical modeling to analyze the relationship between vegetation cover and mental health disorder cases. Specifically, an Enhanced Vegetation Index was used to detect urban vegetation, and Bayesian Poisson lognormal models were implemented to study the association between vegetation and mental health disorder cases of males and females in the 0-19, 20-44, 45-64, and ≥65 years age groups, after controlling for marginalization and unmeasured (latent) spatial and nonspatial covariates at the neighborhood level.ResultsThe results suggest that even after adjusting for marginalization, there were significant age- and sex-specific effects of vegetation on the prevalence of mental health disorders in Toronto. Mental health disorders were negatively associated with the vegetation cover for males aged 0-19 years (−7.009; 95% CI −13.130 to −0.980) and for both males (−4.544; 95% CI −8.224 to −0.895) and females (−3.513; 95% CI −6.289 to −0.681) aged 20-44 years. However, for older adults in the 45-64 and ≥65 years age groups, only the marginalization covariates were significantly associated with mental health disorder cases. In addition, a substantial influence of the unmeasured (latent) and spatially structured covariates was detected in each model (relative contributions>0.7), suggesting that the variations in area-specific relative risk were mainly spatial in nature.ConclusionsAs significant and negative associations between vegetation and mental health disorder cases were found for young males and females, investments in urban greenery can help reduce the future burden of mental health disorders in Canada. The findings highlight the urgent need to understand the age-sex dynamics of the interaction between surrounding vegetation and urban dwellers and its subsequent impact on mental well-being.

Exploring the spatiotemporal correlation among traffic crashes on arterials with signal coordination

oordinated signalization has been applied on arterials since 1960s. However, long-term effects of signal coordination on traffic safety have not been fully discussed, especially in terms of the temporal and spatial variations among crashes. With Bayesian Poisson lognormal model, the paper compares three types of spatiotemporal correlations on the coordinated arterials, including the fixed spatial and linear temporal variation, the fixed spatial and linear-quadratic temporal variation, and the time-varying spatial interaction. With 9-year crash data from Michigan, the study demonstrates that (1) different models should be separately applied for the coordinated arterials, e.g. BPLM with the time-varying spatial interaction fits roadway segments, and BPLM with the fixed spatial and linear temporal variation excels at signalized intersections and (2) the temporal and spatial variations are correlated on roadway segments. The findings emphasize the importance of conducting the safety analysis regularly and dynamically monitoring driving behaviors along the roadway segments.

Investigation of Safety-in-Numbers for Pedestrians and Bicyclists at a Macroscopic Level with Various Exposure Variables

Safety-in-numbers is a phenomenon whereby the crash risks of road users decrease when their numbers increase. Although several previous studies have confirmed safety-in-numbers at a microscopic level (e.g., intersection), few studies have investigated safety-in-numbers at a macroscopic level (or zonal level). In this study, safety-in-numbers is investigated at a larger scale unit, the metropolitan statistical area (MSA), which is usually composed of multiple counties in the U.S.A. Various pedestrian and bicyclist exposure data were obtained from the U.S. National Household Travel Survey (i.e., trips, miles, and hours). A series of Bayesian Poisson lognormal models confirm safety-in-numbers with the different exposure variables at a large-scale geographic level (i.e., MSA). The findings imply that regional travel behavior and cultures of respect for vulnerable road users play a key role in determining the level of pedestrian and bicyclist safety. In addition, the results reveal other factors important to vulnerable road user involved crashes, including but not limited to the climate, demographic, socioeconomic, and travel characteristics of the study regions.

On the safety effects of spatially aggregated traffic flow parameters at macro-levels

Traffic flow parameters have been found to significantly affect crash risk at micro-levels. If such effects do exist at macro-levels, at least two benefits could be expected: (1) the performance and estimates of planning-based crash models could be improved and (2) useful safety knowledge could be provided for regional traffic management. In this article, a flow-based spatial unit was developed by a graph-cut minimization method, based on which regional management strategies are often applied. The graph-cut method partitioned the central area of Kunshan, China, into multiple sub-regions (i.e. graph-cut unit), considering traffic density homogeneity. Bayesian Poisson lognormal models with conditional autoregressive priors were utilized to examine the safety effects of traffic flow parameters, based on the traditional planning-based units and the flow-based graph-cut units. According to the results, no significant traffic flow effect was found for the traffic analysis zone–based model. Traffic flow parameters resulted in a decreased model performance and potential endogeneity issues for the census tract–based model. However, traffic flow effects were found significant for the graph-cut-based model, with an improved model performance. In general, the safety effects of macro-level traffic flow need to be considered for flow-based units developed for regional management.

Analysis and comparison of safety models using average daily, average hourly, and microscopic traffic

There have been plenty of traffic safety studies based on average daily traffic (ADT), average hourly traffic (AHT), or microscopic traffic at 5 min intervals. Nevertheless, not enough research has compared the performance of these three types of safety studies, and seldom of previous studies have intended to find whether the results of one type of study is transferable to the other two studies. First, this study built three models: a Bayesian Poisson-lognormal model to estimate the daily crash frequency using ADT, a Bayesian Poisson-lognormal model to estimate the hourly crash frequency using AHT, and a Bayesian logistic regression model for the real-time safety analysis using microscopic traffic. The model results showed that the crash contributing factors found by different models were comparable but not the same. Four variables, i.e., the logarithm of volume, the standard deviation of speed, the logarithm of segment length, and the existence of diverge segment, were positively significant in the three models. Additionally, weaving segments experienced higher daily and hourly crash frequencies than merge and basic segments. Then, each of the ADT-based, AHT-based, and real-time models was used to estimate safety conditions at different levels: daily and hourly, meanwhile, the real-time model was also used in 5 min intervals. The results uncovered that the ADT- and AHT-based safety models performed similar in predicting daily and hourly crash frequencies, and the real-time safety model was able to provide hourly crash frequency.

Rethinking safety in numbers: are intersections with more crossing pedestrians really safer?

ObjectiveTo advance the interpretation of the ‘safety in numbers’ effect by addressing the following three questions. How should the safety of pedestrians be measured, as the safety of individual pedestrians...

Evaluating Long-Range Regional Safety with Scenario Planning Analysis

This study attempted to evaluate the long-range regional safety of alternative growth scenarios. In doing so, a scenario planning process was undertaken for the midregion of New Mexico. This paper discusses the manifold aspects of data preparation that went into a scenario planning process. The study used land use and travel demand models to parameterize future scenarios and applied these parameters as covariates to model total and severe crashes separately for the future planning horizon with the use of three statistical modeling techniques: a negative binomial model, a Bayesian Poisson–lognormal model without accounting for spatial heterogeneity, and a spatial Bayesian Poisson–lognormal model. These models were investigated and compared for their forecasts for three alternative scenarios that were developed during the scenario planning process. The method adopted for generating and assembling future covariates from the land use and travel demand model accounted for their intricate nature and complexity to be geographically appropriate so that they might reliably synthesize future scenarios. The widest band of increase in crashes was observed from nonspatial Bayesian models. Despite their better predictive fit compared with the negative binomial models, they provided similar or worse safety forecasts for the alternative scenarios. The spatial models provided the least variability in their forecasts for all the alternative scenarios and had the best goodness of fit values. It was concluded that the approach of a safety forecast (short term versus long term) can be dictated by the accuracy of the predictability needed and expected from a regional safety model.

Sensitivity analysis in the context of regional safety modeling: Identifying and assessing the modifiable areal unit problem

A wide array of spatial units has been explored in current regional safety analysis. Since traffic crashes exhibit extreme spatiotemporal heterogeneity which has rarely been a consideration in partitioning these zoning systems, research based on these areal units may be subjected to the modifiable areal unit problem (MAUP).This study attempted to conduct a sensitivity analysis to quantitatively investigate the MAUP effect in the context of regional safety modeling. The emerging regionalization method-RECDAP (regionalization with dynamically constrained agglomerative clustering and partitioning) was employed to aggregate 738 traffic analysis zones in the county of Hillsborough to 14 zoning schemes at an incremental step-size of 50 zones based on spatial homogeneity of crash risk. At each level of aggregation, a Bayesian Poisson lognormal model and a Bayesian spatial model were calibrated to explain observed variations in total/severe crash counts given a number of zone-level factors.Results revealed that as the number of zones increases, the spatial autocorrelation of crash data increases. The Bayesian spatial model outperforms the Bayesian Poisson-lognormal model in accurately accounting for spatial autocorrelation effects, unbiased parameter estimates, and model performance, especially in cases with higher disaggregated levels. Zoning schemes with higher number of zones tend to have increasing number of significant variables, more stable coefficient estimation, smaller standard error, whereas worse model performance. The variables of population density and median household income show consistently significant effects on crash risk and are robust to variation in data aggregation. The MAUP effects may be significantly reduced if we just maintain at about 50% of the original number of zones (350 or larger). The present study highlights MAUP that is generally ignored by transportation safety analysts, and provides insights into the nature of parameter sensitivity to data aggregation in the context of regional safety modeling.

Analysis of residence characteristics of at-fault drivers in traffic crashes

In recent years many studies have investigated traffic crashes with various contributing factors at the macroscopic level. Nevertheless, while previous studies have concentrated only on zones where the crash occurred, there have been few studies that focused on residence characteristics associated with the origin of the drivers causing traffic crashes, so called at-fault drivers. Intuitively, it is reasonable to assume that the number of at-fault drivers is related to socio-demographic features of the at-fault drivers’ residence area. Thus, the main objective of this study is to find out the relationship between the number of at-fault drivers and zonal characteristics of the residence where at-fault drivers came from. The Bayesian Poisson-lognormal model was adopted to find out the contributing factors of the residence zones on the number of crashes based on the at-fault drivers. The findings from the study implied that the crash occurrence is not only affected by roadway/traffic factors but also by several demographic and socioeconomic characteristics of residence zones. The result from this study can be used to identify zones with a higher potential of at-fault drivers; thus we can concentrate on these zones for safety treatments, including more targeted awareness, education or stricter enforcement.

Application of Poisson random effect models for highway network screening

In recent years, Bayesian random effect models that account for the temporal and spatial correlations of crash data became popular in traffic safety research. This study employs random effect Poisson Log-Normal models for crash risk hotspot identification. Both the temporal and spatial correlations of crash data were considered. Potential for Safety Improvement (PSI) were adopted as a measure of the crash risk. Using the fatal and injury crashes that occurred on urban 4-lane divided arterials from 2006 to 2009 in the Central Florida area, the random effect approaches were compared to the traditional Empirical Bayesian (EB) method and the conventional Bayesian Poisson Log-Normal model. A series of method examination tests were conducted to evaluate the performance of different approaches. These tests include the previously developed site consistence test, method consistence test, total rank difference test, and the modified total score test, as well as the newly proposed total safety performance measure difference test. Results show that the Bayesian Poisson model accounting for both temporal and spatial random effects (PTSRE) outperforms the model that with only temporal random effect, and both are superior to the conventional Poisson Log-Normal model (PLN) and the EB model in the fitting of crash data. Additionally, the method evaluation tests indicate that the PTSRE model is significantly superior to the PLN model and the EB model in consistently identifying hotspots during successive time periods. The results suggest that the PTSRE model is a superior alternative for road site crash risk hotspot identification.

Implications of Pedestrian Safety Planning Factors in Areas with Minority and Low-Income Populations

ABSTRACT Minority status and low-income are two important criteria on which environmental-justice areas are being identified. These unique areal characteristics have been of particular interest to traffic safety analysts in investigating pedestrian safety planning factors in deprived areas. In this study, a pedestrian crash model was developed for these areas explicitly accounting for the spatial autocorrelations of different zonal factors commonly used in the traditional transportation planning models. It was found that the Bayesian Poisson-lognormal model with a spatial effect term performed the best compared to the models without accounting for spatial autocorrelations.

Macroscopic spatial analysis of pedestrian and bicycle crashes

This study investigates the effect of spatial correlation using a Bayesian spatial framework to model pedestrian and bicycle crashes in Traffic Analysis Zones (TAZs). Aggregate models for pedestrian and bicycle crashes were estimated as a function of variables related to roadway characteristics, and various demographic and socio-economic factors. It was found that significant differences were present between the predictor sets for pedestrian and bicycle crashes. The Bayesian Poisson-lognormal model accounting for spatial correlation for pedestrian crashes in the TAZs of the study counties retained nine variables significantly different from zero at 95% Bayesian credible interval. These variables were – total roadway length with 35mph posted speed limit, total number of intersections per TAZ, median household income, total number of dwelling units, log of population per square mile of a TAZ, percentage of households with non-retired workers but zero auto, percentage of households with non-retired workers and one auto, long term parking cost, and log of total number of employment in a TAZ. A separate distinct set of predictors were found for the bicycle crash model. In all cases the Bayesian models with spatial correlation performed better than the models that did not account for spatial correlation among TAZs. This finding implies that spatial correlation should be considered while modeling pedestrian and bicycle crashes at the aggregate or macro-level.