Between Dec 19, 2024 and Jan 2, 2025, datasets can be submitted to DRUM but will not be processed until after the break. Staff will not be available to answer email during this period, and will not be able to provide DOIs until after Jan 2. If you are in need of a DOI during this period, consider Dryad or OpenICPSR. Submission responses to the UDC may also be delayed during this time.

Browsing by Author "Ermagun, Alireza"

Now showing 1 - 18 of 18
  • Thumbnail Image
    Item
    Accessibility and Transit Performance
    (2015) Ermagun, Alireza; Levinson, David, M
    This study disentangles the impact of financial and physical dimensions of transit service operators on net transit accessibility for 46 of the 50 largest metropolitan areas in the United States. To investigate this interaction along with the production efficiency of transit agencies, two types of analysis are used: a set of linear and quadratic regressions and a data envelopment analysis. We find that vehicle revenue kilometers and operational expenses play a pivotal role in enhancing the accessibility to jobs by transit. The bivariate linear regression models indicate a 1% increase in operational expenses and vehicle revenue kilometers increase the number of jobs that can be reached within 30 minutes by 0.96 and 0.95%, respectively. The results of the quadratic functional form, also, show transit services may have both increasing and decreasing accessibility returns to scale depending on system size, and the results are sensitive to the model used. Overall, the highest system efficiency (access produced per input) is found in the New York, Washington, and Milwaukee metropolitan areas, while Riverside, Detroit, and Austin perform with the lowest efficiency.
  • Thumbnail Image
    Item
    Accessibility-based Evaluation of Transit Projects
    (2016-08-01) Palmateer, Chelsey; Owen, Andrew; Levinson, David M; Ermagun, Alireza
    This study uses the accessibility-based evaluation method to unpack the interaction effect of transit oriented development and a new transit hub, using the San Francisco Transbay Transit Center Development Plan project. We reveal both the transit oriented development and transit changes positively affect accessibility to jobs and accessibility to workers. However, the magnitude of effects for the transit changes alone are minimal in comparison to the effects of the anticipated transit oriented development changes. This indicates that in areas where there already is transit service, the development of land near the transit service can have a greater impact on accessibility levels than the improvement of connections between transit services. We also unravel the increase in accessibility at the project-level and determine that the increase is greater than the sum of the contributions of the individual portions of the project. This demonstrates that transit changes and transit oriented development have a superadditive effect, although it is negligible in our case.
  • Thumbnail Image
    Item
    Analyses of Bicycle and Pedestrian Trail Traffic: New Tools for Modeling User Expenditures and Demand
    (2017-08) Ermagun, Alireza
    Despite the importance of multi-use trails in urban non-motorized transportation networks, transportation planners, engineers, and trail managers lack tools for describing economic activity associated with local trail use and for predicting bicycle and pedestrian demand for trails. New tools are needed to plan and prioritize investments in new facilities and to inform management and maintenance of trail infrastructure. Among other needs, they need tools to predict (1) expenditures by local users to support local economic development initiatives and assess neighborhood effects of proposals for trail development and (2) trail traffic demand for optimizing investments and managing maintenance of systems and facilities. This thesis responds to these needs and augments the burgeoning literature on trail traffic analysis by developing models of trail-related expenditures and mode-specific trail demand models. From the expenditures by local users side, using the results of intercept surveys completed by 1,282 trail users on the Central Ohio Greenway trail network in 2014, this thesis estimates the probabilities and patterns that different types of trail users will make expenditures. Approximately one-fifth of trail users reported spending between US$15 and US$20 for food, drink, and other incidental items. Across all trail users the average expenditure by individuals is about US$3 per visit. All else equal, cyclists are more than twice as likely than other users to report expenditures. Users visiting trails principally for recreation are 53% more likely to spend, while users visiting trails mainly for exercise were about 19% less likely. Both longer trips to and on the trails are associated with higher spending. From the trail traffic demand side, this thesis employs trail traffic volumes recorded at 15-minute intervals for 32 multi-use trails located in 13 urban areas across the United States from January 1, 2014 through February 16, 2016. The results of analyses indicate (1) daily trail traffic varies substantially – over three orders of magnitude – across the monitoring stations included in the study; (2) daily trail traffic is highly correlated with weather, and the parabola form of weather parameters works well for modeling variables such as temperature, where trail use is associated with warmer temperatures, but only up to a point at which higher temperatures then decrease use; (3) bicyclists and pedestrians respond differently to variations in weather, and their responses vary both within and across regions; (4) with only a few exceptions, average daily pedestrians (ADP) and average daily bicyclists (ADB) are correlated with different variables, and the magnitude of effects of variables that are the same varies significantly between the two modes; (5) the mean relative percentage error (MRPE) for bicyclist, pedestrian, and mixed-mode demand models, respectively, are 65.4%, 85.3%, and 45.9%; (6) although using multimodal monitoring networks enables us to juxtapose the bicyclist demand with pedestrian demand, there is not a significant improvement in predicting total demand using multimodal sensors; (7) a new post-validation procedure improves the demand models, reducing the MRPE of bicyclist, pedestrian, and mixed-mode models by 27.2%, 32.1%, and 14.1%. Transportation planners, engineers, and trail managers can use these results to estimate the effects of weather and climate on trail traffic and to plan and manage facilities more effectively. The developed models also can be used in practical applications such as selection of route corridors and prioritization of investments where order-of-magnitude estimates suffice.
  • Thumbnail Image
    Item
    Development and Application of the Network Weight Matrix to Predict Traffic Flow for Congested and Uncongested Conditions
    (2016-08-01) Ermagun, Alireza; Levinson, David M
    To capture a more realistic spatial dependence between traffic links, we introduce two distinct network weight matrices to replace spatial weight matrices used in traffic forecasting methods. The first stands on the notion of betweenness centrality and link vulnerability in traffic networks. To derive this matrix, we assume all traffic flow is assigned to the shortest path, and thereby we used Dijkstra's algorithm to find the shortest path. The other relies on flow rate change in traffic links. For forming this matrix, we employed user equilibrium assignment and the method of successive averages (MSA) algorithm to solve the network. The components of the network weight matrices are a function not simply of adjacency, but of network topology, network structure, and demand configuration. We tested and compared the network weight matrices in different traffic conditions using Nguyen-Dupuis network. The results led to a clear and unshakable conclusion that spatial weight matrices are unable to capture the realistic spatial dependence between traffic links in a network. Not only do they overlook the competitive nature of traffic links, but they also ignore the role of network topology and demand configuration. In contrast, the flow-weighted betweenness method significantly operates better than unweighted betweenness to measure realistic spatial dependence between traffic links, particularly in congested traffic conditions. The results disclosed that this superiority is more than 2 times in congested flow situations. However, forming this matrix requires considerable computational effort and information. If the network is uncongested the network weight matrix stemming from betweenness centrality is sufficient.
  • Thumbnail Image
    Item
    The Influences of the Hiawatha LRT on Changes in Travel Behavior: A Retrospective Study on Movers
    (Center for Transportation Studies, University of Minnesota, 2015-11) Cao, Jason; Ermagun, Alireza
    Following scant evidence for the effects of proximity to rail transit on auto use, we pinpoint the impacts of rail transit and neighborhood characteristics on both transit and car use in the Minneapolis-St. Paul metropolitan area. In this vein, we apply the structural equations modeling approach on 597 residents who moved into the Hiawatha light rail transit (LRT) corridor after it opened. Using a quasi-longitudinal design to compare the behavior of movers into the Hiawatha and control corridors, we found that the Hiawatha LRT acts as both a catalyst and a magnet. Movers into the Hiawatha corridor experience transit improvement, which increases transit use and reduces car use. The LRT also enables transit-liking people who were unable to realize their preference previously to relocate near the LRT. However, the LRT has no significant effects on changes in auto ownership. This suggests that besides transit infrastructure, planners should promote transit-friendly neighborhood characteristics.
  • Thumbnail Image
    Item
    Intra-household Bargaining for School Trip Accompaniment of Children: A Group Decision Approach
    (2016-08-01) Ermagun, Alireza; Levinson, David M
    This paper tests a group decision-making model to examine the school travel behavior of students 6-18 years old in the Minneapolis-St. Paul Metropolitan area. The school trip information of 1,737 two-parent families with a student is extracted from Travel Behavior Inventory data collected by the Metropolitan Council between the Fall 2010 and Spring 2012. The proposed model has four distinct characteristics including: (1) considering the student explicitly in the model, (2) allowing for bargaining or negotiation within households, (3) quantifying the intra-household interaction among family members, and (4) determining the decision weight function for household members. This framework also covers a household with three members, namely, a father, a mother, and a student, and unlike other studies it is not limited to dual-worker families. To test the hypotheses we developed two models, each with and without the group-decision approach. The models are separately developed for different age groups, namely students 6-12 and 12-18 years old. This study considered a wide range of variables such as work status of parents, age and gender of students, mode of travel, and distance to school. The findings of this study demonstrate that the elasticities of the two modeling approaches differ not only in the value, but in the sign in some cases. In 63 percent of the cases the unitary household model underestimates the results. More precisely, the elasticities of the unitary household model are as large as 2 times more than that of the group-decision model in 20 percent of cases. This is a direct consequence of model misspecification that misleads both long- and short-term policies where the intra-household bargaining and interaction is overlooked in travel behavior models.
  • Thumbnail Image
    Item
    An Introduction to the Network Weight Matrix
    (2016-08-01) Ermagun, Alireza
    This study introduces the network weight matrix as a replacement for the spatial weight matrix to measure the spatial dependence between links of a network. This matrix stems from the concept of betweenness centrality and vulnerability in network science. The elements of the matrix are a function not simply of proximity, but of network topology, network structure, and demand configuration. The network weight matrix has distinctive characteristic, which are capable of reflecting spatial dependence between traffic links: (1) The elements are allowed to have negative and positive values, which capture competitive and complementary nature of links, (2) The diagonal elements are not fixed to zero, which takes the self-dependence of a link upon itself into consideration, and (3) The elements not only reflect the spatial dependence based on the network structure, but they acknowledge the demand configuration as well. We verified the network weight matrix by modeling traffic flows in a 3x3 grid test network with 9 nodes and 24 directed links connecting 72 origin-destination (OD) pairs. The results disclose models encompassing the network weight matrix outperform both models without spatial components and models with the spatial weight matrix. This leads inexorably to the conclusion that the network weight matrix represents a more accurate and defensible spatial dependency between traffic links, and thereby augments traffic flow prediction.
  • Thumbnail Image
    Item
    Network Econometrics and Traffic Flow Analysis
    (2016-10) Ermagun, Alireza
    This dissertation introduces concepts, theories, and methods dealing with network econometrics to gain a deeper understanding of how the components interact in a complex network. More precisely, it introduces distinctive network weight matrices to extract the existing spatial dependency between traffic links. The network weight matrices stem from the concepts of betweenness centrality and vulnerability in network science. Their elements are a function not simply of proximity, but of network topology, network structure, and demand configuration. The network weight matrices are tested in congested and uncongested traffic conditions in both simulation-based and real-world environments. The results of the analysis lead to a conclusion that traditional spatial weight matrices are unable to capture the realistic spatial dependency between traffic links in a network. Not only do they overlook the competitive nature of traffic links, but they also ignore the role of network topology and demand configuration in measuring the spatial dependence between traffic links. However, the proposed network weight matrices substitute for traditional spatial weight matrices and exhibit the capability to overcome these deficiencies. The network weight matrices are theoretically defensible in account of acknowledging traffic theory. They capture the competitive and complementary nature of links and embed additional network dynamics such as cost of links and demand configuration. Building on real-world data analysis, the results contribute to the conclusion that in a network comprising links in parallel and series, both negative and positive correlations show up between links. The strength of the correlation varies by time-of-day and day-of-week. Strong negative correlations are observed in rush hours, when congestion affects travel behavior. This correlation occurs mostly in parallel links, and in far upstream links where travelers receive information about congestion (for instance from media, variable message signs, or personal observations of propagating shockwaves) and are able to switch to substitute paths. Irrespective of time-of-day and day-of-week, a strong positive correlation is observed between upstream and downstream sections. This correlation is stronger in uncongested regimes, as traffic flow passes through the consecutive links in a shorter time and there is no congestion effect to shift or stall traffic.
  • Thumbnail Image
    Item
    Physical Activity in School Travel: A Cross-Nested Logit Approach
    (2016) Ermagun, Alireza; Levinson, David, M
    This paper considers school access by both active (walk, bike), quasi-active (walk to transit) and non-active modes (car) in a two-level cross-nested logit framework. A sample of 3,272 middle and high school students was collected in Tehran. The results of the cross-nested logit model suggest that for people who choose walking, increasing a 1 percent in home-to-school distance reduces the probability of walking by 3.51 percent. While, this reduction is equal to 2.82 and 2.27 percent as per the multinomial and nested logit models, respectively. This is a direct consequence of the model specification that results in underestimating the effect of distance by 1.24 percent. It is also worth mentioning that, a one percent increase in home-to-school distance diminishes the probability of taking public transit by 1.04 among public transit users, while increases the probability of shifting to public transit from walking by 1.39 percent. Further, a one percent increase of the distance to public transport, decreases the probability of students' physical activity, approximately, 0.04 percent.
  • Thumbnail Image
    Item
    Safety in Numbers and Safety in Congestion for Bicyclists and Motorists at Urban Intersections
    (2017) Carlson, Kristin; Ermagun, Alireza; Murphy, Brendan; Owen, Andrew; Levinson, David M.
    This study assesses the estimated crashes per bicyclist and per vehicle as a function of bicyclist and vehicle traffic, and tests whether greater traffic reduces the per-car crash rate. We present a framework for comprehensive bicyclist risk assessment modeling, using estimated bicyclist flow per intersection, observed vehicle flow, and crash records. Using a two-part model of crashes, we reveal that both the annual average daily traffic and daily bicyclist traffic have a diminishing return to scale in crashes. This accentuates the positive role of safety in numbers. Increasing the number of vehicles and cyclists decelerates not only the probability of crashes, but the number of crashes as well. Measuring the elasticity of the variables, it is found that a 1% increase in the annual average daily motor vehicle traffic increases the probability of crashes by 0.14% and the number of crashes by 0.80%. However, a 1% increase in the average daily bicyclist traffic increases the probability of crashes by 0.09% and the number of crashes by 0.50%. The saturation point of the safety in numbers for bicyclists is notably less than for motor vehicles. Extracting the vertex point of the parabola functions examines that the number of crashes starts decreasing when daily vehicle and bicyclist traffic per intersection exceed 29,568 and 1,532, respectively.
  • Thumbnail Image
    Item
    Safety in Numbers: Pedestrian and Bicyclist Activity and Safety in Minneapolis
    (Center for Transportation Studies, University of Minnesota, 2018-03) Carlson, Kristin; Murphy, Brendan; Ermagun, Alireza; Levinson, David; Owen, Andrew
    This investigation aims to evaluate whether the Safety in Numbers phenomenon is observable in the midwestern U.S. city of Minneapolis, Minnesota. Safety in Numbers (SIN) refers to the phenomenon that pedestrian safety is positively correlated with increased pedestrian traffic in a given area. Walking and bicycling are increasingly becoming important transportation modes in modern cities. Proper placement of non-motorized facilities and improvements has implications for safety, accessibility, and mode choice, but proper information regarding estimated non-motorized traffic levels is needed to locate areas where investments can have the greatest impact. Assessment of collision risk between automobiles and non-motorized travelers offers a tool that can help inform investments to improve non-motorized traveler safety. Models of non-motorized crash risk typically require detailed historical multimodal crash and traffic volume data, but many cities do not have dense datasets of non-motorized transport flow levels. Methods of estimating pedestrian and bicycle behavior that do not rely heavily on high-resolution count data are applied in this study. Pedestrian and cyclist traffic counts, average automobile traffic, and crash data from the city of Minneapolis are used to build models of crash frequencies at the intersection level as a function of modal traffic inputs. These models determine whether the SIN effect is observable within the available datasets for pedestrians, cyclists, and cars, as well as determine specific locations within Minneapolis where non-motorized travelers experience elevated levels of risk of crashes with automobiles.
  • Thumbnail Image
    Item
    Spatiotemporal Short-term Traffic Forecasting using the Network Weight Matrix and Systematic Detrending
    (2017-08) Ermagun, Alireza; Levinson, David M
    This study examines the dependency between traffic links using a three-dimensional data detrend- ing algorithm to build a network weight matrix in a real-world example. The network weight matrix reveals how links are spatially dependent in a complex network and detects the competi- tive and complementary nature of traffic links. We model the traffic flow of 140 traffic links in a sub-network of the Minneapolis - St. Paul highway system for both rush hour and non-rush hour time intervals, and validate the extracted network weight matrix. The results of the modeling indi- cate: (1) the spatial weight matrix is unstable over time-of-day, while the network weight matrix is robust in all cases and (2) the performance of the network weight matrix in non-rush hour traffic regimes is significantly better than rush hour traffic regimes. The results of the validation show the network weight matrix outperforms the traditional way of capturing spatial dependency between traffic links. Averaging over all traffic links and time, this superiority is about 13.2% in rush hour and 15.3% in non-rush hour, when only the 1st -order neighboring links are embedded in modeling. Aside from the superiority in forecasting, a remarkable capability of the network weight matrix is its stability and robustness over time, which is not observed in spatial weight matrix. In addition, this study proposes a naïve two-step algorithm to search and identify the best look-back time win- dow for upstream links. We indicate the best look-back time window depends on the travel time between two study detectors, and it varies by time-of-day and traffic link.
  • Thumbnail Image
    Item
    Spatiotemporal Traffic Forecasting: Review and Proposed Directions
    (2016-08-01) Ermagun, Alireza; Levinson, David M
    This paper systematically reviews studies that forecast short-term traffic conditions using spatial dependence between links. We synthesize 130 extracted research papers from two perspectives: (1) methodological framework, and (2) approach for capturing and incorporating spatial information. From the methodology side, spatial information boosts the accuracy of prediction, particularly in congested traffic regimes and for longer horizons. There is a broad and longstanding agreement that non-parametric methods outperform the naive statistical methods such as historical average, real time profile, and exponential smoothing. However, to make an inexorable conclusion regarding the performance of neural network methods against STARIMA family models, more research is needed in this field. From the spatial dependency detection side, we believe that a large gulf exists between the realistic spatial dependence of traffic links on a real network and the studied networks. This systematic review highlights that the field is approaching its maturity, while it is still as crude as it is perplexing. It is perplexing in the conceptual methodology, and it is crude in the capture of spatial information.
  • Thumbnail Image
    Item
    The Role of Transit Service Area Definition in Accessibility-based Evaluation
    (2016-08-01) Palmateer, Chelsey; Owen, Andrew; Levinson, David M; Ermagun, Alireza
    This empirical study examines the importance of service area definition, when utilizing accessibility-based evaluation in transit projects. We analyze two transit projects: (1) Metro Transit A-Line in Minnesota and (2) Harris County Transit Re-Imagined Bus Network in Texas. The results indicate that the choice of transit service areas have a significant impact on the value of ab- solute accessibility measures. The trend shows the narrower the service area, the higher the value of the absolute accessibility measure. The results, however, are inconsistent between projects when relative accessibility measures such as percentage change between scenarios is used as an accessibility-based evaluation measure. We conclude service area definition is of only moderate importance for scenario comparisons within the same analysis boundary. When comparing different regions or areas within different boundaries, the service area definition could have a significant impact on all results. This is case-dependent and varies greatly from project to project, which requires calculating both the absolute and relative accessibility measures in an accessibility-based evaluation. In addition, decomposing the accessibility changes in the separate portions of transit projects reveals that the light rail investments have negligible impacts on accessibility levels, while restructuring of the bus network has a slight positive impact on accessibility levels. The findings have important implications for the deployment of accessibility-based evaluation on transit projects.
  • Thumbnail Image
    Item
    Traffic Flow Variation and Network Structure
    (2017) Ermagun, Alireza; Levinson, David M.
    This study defines and detects competitive and complementary links in a complex network and constructs theories illustrating how the variation of traffic flow is interconnected with network structure. To test the hypotheses, we extract a grid-like sub-network containing 140 traffic links from the Minneapolis - St. Paul highway system. We reveal a real-world traffic network comprises both competitive and complementary links, and there is a negative network dependency between a competitive link pair and a positive network dependency between a complementary link pair. We validate a robust linear relationship between standard deviation of flow in a link and its number of competitive links, its link correlation with competitive links, and its network dependency with both competitive and complementary links. The results indicate the number of competitive links in a traffic network is negatively correlated with the variation of traffic flow in congested regimes as drivers are able to take alternative paths. The results also signify that the more the traffic flow of a link is correlated to the traffic flow of its competitive links, the more the flow variation is in the link. Considering the network dependency, however, it is corroborated that the more the network dependency between a link and its competitive links, the more the flow variation in the link. This is also true for complementary links.
  • Thumbnail Image
    Item
    Traffic Impacts of Bicycle Facilities
    (Minnesota Department of Transportation, 2017-06) Hourdos, John; Lehrke, Derek; Duhn, Melissa; Ermagun, Alireza; Singer-Berk, Lila; Lindsey, Greg
    Engineers need information about interactions between vehicles and bicyclists to design efficient, safe transportation systems. This study involved a review of design guidelines for bicycle facilities, observation of bicycle-vehicle interactions at nine roadways with different types of bicycle facilities, analysis of results, and description of design implications. Facilities observed included buffered and striped bicycle lanes, sharrows, signed shared lanes, and shoulders of various widths. Driver behaviors were categorized as no change in trajectory, deviation within lane, encroachment into adjacent lane, completion of a passing maneuver, and queuing behind cyclists. Drivers on roadways with bicycle lanes were less likely to encroach into adjacent lanes, pass, or queue when interacting with cyclists than drivers on roadways with sharrows, signs designating shared lanes, or no bicycle facilities. Queueing behind cyclists, the most significant impact on vehicular traffic flows, generally was highest on roads with no facilities or shared facilities without marked lanes. Statistical modeling confirmed the descriptive results. Given an objective of increasing predictability of driver behavior, buffered or striped bicycle lanes offer advantages over other facilities. Sharrows may alert drivers to the presence of cyclists, but traffic impacts on roadways with sharrows may not differ significantly from roadways with no facilities. Signs indicating bicyclists may occupy lanes also may alert drivers to the presence of cyclists, but this study provided no evidence that interactions on roadways marked only with signs differ from roadways with no facilities. From the perspective of reducing potential traffic impacts, bicycle lanes are to be preferred over sharrows or signage.
  • Thumbnail Image
    Item
    "Transit Makes you Short": On Health Impact Assessment of Transportation and the Built Environment
    (2015) Ermagun, Alireza; Levinson, David M
    The current research provides a test framework to understand whether and to what extent increasing public transit use and accessibility by transit affect health. To this end, the effect of transit mode share and accessibility by transit on general health, body mass index, and height are investigated, while controlling for socioeconomic, demographic, and physical activity factors. The coefficient-p-value-sample-size chart is created and effect size analysis are conducted to explore whether the transit use is practically significant. Building on the results of the analysis, we found that the transit mode share and accessibility by transit are not practically significant, and the power of large-sample misrepresents the effect of transit on public health. The results, also, highlight the importance of data and variable selection by portraying a significant correlation between transit use and height in a multivariate regression analysis. What becomes clear from this study is that in spite of the mushrooming interdisciplinary studies in the nexus of transportation and health arena, researchers often propose short- and long-term policies blindly, while failing to report the inherent explanatory power of variables. We show that there is a thin line between false positive and true negative results. From the weakness of p-values perspective, further, we strove to alert both researchers and practitioners to the dangerous pitfall deriving from the power of large- samples. Building the results on just significance and sign of the parameter of interest is worthless, unless the magnitude of effect size is carefully quantified post analysis.
  • Thumbnail Image
    Item
    Using Temporal Detrending to Observe the Spatial Correlation of Traffic
    (2016-08-01) Ermagun, Alireza; Levinson, David M; Chatterjee, Snigdhansu
    This empirical study sheds light on the correlation of traffic links under different traffic regimes. We mimic the behavior of real traffic by pinpointing the correlation between 140 freeway traffic links in a sub-network of the Minneapolis - St. Paul highway system with a grid-like network topology. This topology enables us to juxtapose positive correlation with negative correlation, which has been overlooked in short-term traffic forecasting models. To accurately and reliably measure the correlation between traffic links, we develop an algorithm that eliminates temporal trends in three dimensions: (1) hourly dimension, (2) weekly dimension, and (3) system dimension for each link. The correlation of traffic links exhibits a stronger negative correlation in rush hours, when congestion affects route choice. Although this correlation occurs mostly in parallel links, it is also observed upstream, where travelers receive information and are able to switch to substitute paths. Irrespective to the time-of-day and day-of-week, a strong positive correlation is witnessed between upstream and downstream links. This correlation is stronger in uncongested regimes, as traffic flow passes through consecutive links more quickly and there is no congestion effect to shift or stall traffic. The extracted correlation structure can augment the accuracy of short-term traffic forecasting models.