Estimating Effects of Passenger Dwell and Non-Passenger Delay on Overall Bus Trip Time: A Hierarchical Modeling Approach
For local bus service, in-vehicle travel time is composed of time spent in motion and time stopped. Time in motion is dependent on factors common to general traffic travelling on a road segment, while time stopped has some features in common with general traffic (i.e., traffic signals) and some unique to buses such as passenger dwell time. In addition to passenger dwell, there are multiple sources of delay from serving a bus stop, such as deceleration, acceleration, and signal delay. To improve overall travel time, transit agencies can attempt to reduce dwell and delay, but the relative improvements resulted will depend on the contribution of the various components of the bus “time budget.” This paper attempts to quantify the contribution of passenger dwell time, non-passenger delay and in-motion time to total travel time. Using a stop-level data source, passenger dwell time and non-passenger delay are calculated per stop within each trip. A hierarchical probabilistic model is then used to estimate the effects of these components on overall travel time. We demonstrate that for a local, high-frequency route, the fraction of total travel time that a bus spends on stop-associated non-passenger delay is about twice as much as what it spends on passenger dwell. Additionally, the contribution of passenger dwell time is additive to total travel time, whereas non-passenger delay around bus stops is multiplicative. Thus, when transit agencies consider techniques to improve travel time with limited resources, the modeling approach can suggest prioritization of efforts.
Public transit rider satisfaction is well-studied in the academic literature and transit industry. Numerous studies have focused on the factors that drive overall satisfaction and thus provide ample insights to transit agencies on investment priorities. However, there is less published research on the difference in satisfaction across transit mode (light rail, commuter rail, bus), bus route-type (express, arterial bus rapid transit, local service), or demographic groups. This study builds the body of research by providing a comprehensive assessment of public transit rider satisfaction among Metro Transit riders in the Minneapolis/St. Paul metropolitan area. Additionally, it proposes a methodology for analyzing surveys that addresses the categorical and interdependent nature of survey data – a process that employs Gower’s distance and a partitioning around medoids (PAM) clustering algorithm to segment riders based on attitudes along with a Bayesian logistic regression model to profile the unique identified clusters. Light rail, arterial bus rapid transit, express, and particularly commuter rail riders were much more likely to be satisfied when compared to local bus riders. Satisfaction tended to increase with age, low and high-income riders were more satisfied than middle income riders, people of color tended to have slightly lower satisfaction than white riders, while riders who reported having a disability were somewhat more satisfied. Transit reliant riders tended to be less satisfied, whereas new transit riders (less than two years of riding experience) were more satisfied than more experienced riders. Riders who had experienced various forms of street harassment on transit were less satisfied.
Responsible bus accidents—accidents in which an operator’s actions largely explain why an accident occurred—decrease service reliability and safety, cause property damage, and lead to injury or even death. These effects make the reduction of avoidable accidents a high priority for transit authorities. This study used innovative techniques to identify factors that affected the probability of a responsible bus accident in the Minneapolis–Saint Paul, Minnesota, metropolitan area. The study examined data on the trip level and used the random forest model to identify the factors that most strongly affected the likelihood of an accident. In addition, this study used the partial dependency function of the random forest model as an aid in identifying nonlinear relationships between predictor and response variables. This function informs variable transformation in a traditional logistic model, which accurately classifies test set trips as accidents and nonaccidents at a rate of about 72%. Many results fell in line with past literature, but new insights were gained. Specifically, in keeping with most past studies, accidents were more likely when the bus operator was older, female, fatigued, inexperienced, driving a larger bus, or driving on a dense urban route. One new insight of this study is that bus drivers are at greater risk toward the middle of their shift, especially when traffic is dense. Bus drivers’ risk greatly increases if they (a) did not work the previous day and (b) worked longer hours the previous week. Because previous studies have not examined accidents on the trip level, they have not found these more specific work-hour relationships that capture elements of alertness, fatigue, and complacency.
What determines the likelihood a resident will take public transit? To understand the relative effects of various determinants of ridership, this study focuses on the effects of gasoline prices, controlling for other determinants such as number of workdays in a given month, traffic, unemployment rate, and service quality. As gasoline prices increase, it becomes more expensive to drive a car. Thus, customers would likely shy away from driving and substitute towards the alternatives; one of which is public transportation. This study aims to find the relationship between gasoline prices and bus ridership at a disaggregated level (route type level) in the Twin Cities using Ordinary Least Squares and Autoregressive Integrated Moving Average estimators. The crossprice elasticities of local and express bus ridership with respect to gasoline prices are 0.139 and 0.220, respectively.
Have you wondered why some countries have very different income levels? For example, what is the reason behind Cambodia being one of the poorest countries in the world and the United States is on the other end of the spectrum? In this paper, I model the response variable as levels of income classified by the World Bank which includes Low, Lower Middle, Upper Middle, and High Income. My main predictor of interest is constitutional form which includes Monarchy, Parliamentary, Presidential, and Socialist. I also consider other predictors such as cost of export, agricultural land, population density, whether or not a country is landlocked, and foreign direct investment. Only constitutional form and cost of export appear to be significant predictors of a nation’s income level. I use multinomial logistic regression implemented in nnet package in R. More than half of the coefficients for the constitutional form variable are not statistically significant at 5% significance level so I am reluctant in concluding that constitutional form affects a nation’s income level. Cost of export appears to be more statistically significant. Higher cost of export is associated with a higher probability of a country being in a low income group.