Capturing tour mode and activity choice interdependencies: A co-evolutionary logit modelling approach

It is generally assumed that the choice of transport mode and the choice of including intermediate activities on a work tour are interrelated, but little is known about the nature of the causal relationship. To shed light on this, this paper addresses the question of whether transport mode choice is dependent on the activity choice or vice-versa. A new methodology, referred to as the co-evolutionary approach, is combined with a set of MNL models, one for each choice facet involved, to derive an indication of the order of decisions on an individual level. The models are estimated based on the work tours of a large sample of individuals in the Netherlands. The results suggest that there is substantial variation in the order of the transport mode and activity decisions. However, in the majority of cases the activity decision is made before the mode decision, suggesting that the transport mode and, in particular, the choice between car and public transport is most often ‘adjusted’ to the choice of trip chaining rather than the other way round.     

Modelling daily activity program generation considering within-day and day-to-day dynamics in activity-travel behaviour

This paper presents a comprehensive econometric modelling framework for daily activity program generation. It is for day-specific activity program generations of a week-long time span. Activity types considered are 15 generic categories of non-skeletal and flexible activities. Under the daily time budget and non-negativity of participation rate constraints, the models predict optimal sets of frequencies of the activities under consideration (given the average duration of each activity type). The daily time budget considers at-home basic needs and night sleep activities together as a composite activity. The concept of composite activity ensures the dynamics and continuity of time allocation and activity/travel behaviour by encapsulating altogether the activity types that are not of our direct interest in travel demand modelling. Workers’ total working hours (skeletal activity and not a part of the non-skeletal activity time budget) are considered as a variable in the models to accommodate the scheduling effects inside the generation model of non-skeletal activities. Incorporation of previous day’s total executed activities as variables introduces day-to-day dynamics into the activity program generation models. The possibility of zero frequency of any specific activity under consideration is ensured by the Kuhn-Tucker optimality conditions used for formulating the model structure. Models use the concept of random utility maximization approach to derive activity program set. Estimations of the empirical models are done using the 2002–2003 CHASE survey data set collected in Toronto.

Socio-demographic and built environment influences on the odds of being overweight or obese: The Atlanta experience

This paper explores the association of socio-demographic and built environment characteristics on the odds of being overweight and obese using data from the Atlanta SMARTRAQ travel survey. A new methodological framework based on a multinomial logit (MNL) model and an enhanced odds ratio plot is presented. The use of an MNL model overcomes limitations of many prior studies that employ a sequence of binary logit models to examine multiple weight categories. The use of an enhanced odds ratio plot provides important information into the relative importance of socio-demographic and built environment characteristics. Several new findings for the Atlanta area result from this study. Socio-demographic variables, including age and educational attainment, exhibit a non-linear relationship with the odds of being overweight or obese. Gender, age, ethnicity, and educational attainment are strongly associated with the odds of being overweight or obese, while income and number of students between 5 and 16 years old in the household have smaller effects. Built environment characteristics such as increased net residential densities and enhanced street connectivity are associated with reductions in the odds of being overweight and/or obese. Relative to socio-demographic variables, however, such built environment characteristics have a much smaller impact on describing the odds of being overweight or obese.     

Vehicle classification from low-frequency GPS data with recurrent neural networks

The categorization of the type of vehicles on a road network is typically achieved using external sensors, like weight sensors, or from images captured by surveillance cameras. In this paper, we leverage the nowadays widespread adoption of Global Positioning System (GPS) trackers and investigate the use of sequences of GPS points to recognize the type of vehicle producing them (namely, small-duty, medium-duty and heavy-duty vehicles). The few works which already exploited GPS data for vehicle classification rely on hand-crafted features and traditional machine learning algorithms like Support Vector Machines. In this work, we study how performance can be improved by deploying deep learning methods, which are recently achieving state of the art results in the classification of signals from various domains. In particular, we propose an approach based on Long Short-Term Memory (LSTM) recurrent neural networks that are able to learn effective hierarchical and stateful representations for temporal sequences. We provide several insights on what the network learns when trained with GPS data and contextual information, and report experiments on a very large dataset of GPS tracks, where we show how the proposed model significantly improves upon state-of-the-art results.     

Exploring the links between personal networks, time use, and the spatial distribution of social contacts

Although the study of the role of the social context in travel behavior and activity patterns has recently gained attention, the empirical evidence supporting the relationship between social networks and the temporal and spatial characteristics of social activities is still limited. With this motivation, this paper studies the link between “longer term” (social networks) and “shorter term” (social activities) social decisions, by exploring the intertwined relationship between the individuals’ personal networks attributes, and the spatiotemporal characteristics of their daily social activities. The paper contributes to the literature by adding two key aspects to the study of the role of social networks on travel behavior: the social networks’ structure, and the spatiality of all individuals participating on the social activities. Based on data which link people’s personal networks and time use, and using a structural equation modeling approach, the paper studies the influence of individual and interactional attributes on the duration, distance, and number of people involved in social daily activities. The results show that aspects such as tie social closeness, gender and age similarity, and network density, help to understand social activity duration and distance, complementing traditional socio-demographic aspects such as income, occupation, and accessibility to services. In this way, socio-demographic attributes are not enough to explain the spatiotemporal dimension of daily activities which makes necessary to include variables related to the social context to explain with a higher level of accuracy both the duration and distance traveled to the activity.     

Machine learning applications in activity-travel behaviour research: a review

ABSTRACT

This paper reviews the activity-travel behaviour literature that employs Machine Learning (ML) techniques for empirical analysis and modelling. Machine Learning algorithms, which attempt to build intelligence utilizing the availability of large amounts of data, have emerged as powerful tools in the fields of pattern recognition and big data analysis. These techniques have been applied in activity-travel behaviour studies since the early ’90s when Artificial Neural Networks (ANN) were employed to model mode choice decisions. AMOS, an activity-based modelling system developed in the mid-’90s, has ANN at its core to model and predict individual responses to travel demand management measures. In the dawn of 2000, ALBATROSS, a comprehensive activity-based travel demand modelling system, was proposed by Arentze and Timmermans using Decision Trees. Since then researchers have been exploring ML techniques like Support Vector Machines (SVM), Decision Trees (DT), Neural Networks (NN), Bayes Classifiers, and more recently, Ensemble Learners to model and predict activity-travel behaviour. A large number of publications over the years and an upward trend in the number of published articles over time indicate that Machine Learning is a promising tool for activity-travel behaviour analysis and prediction. This article, first of its kind in the literature, reviews these studies and explores the trends in activity-travel behaviour research that apply ML techniques. The review finds that mode choice decisions have received wide attention in the literature on ML applications. It was observed that most of the studies identify the lack of interpretability as a serious shortcoming in ML techniques. However, very few studies have attempted to improve the interpretability of the models. Further, some studies report the importance of feature engineering in ML-based studies, but very few studies adopt feature engineering before model development. Spatiotemporal transferability of models is another issue that has received minimal attention in the literature. In the end, the paper discusses possible directions for future research in the area of activity-travel behaviour modelling using ML techniques.     

Inferring patterns in the multi-week activity sequences of public transport users

The public transport networks of dense cities such as London serve passengers with widely different travel patterns. In line with the diverse lives of urban dwellers, activities and journeys are combined within days and across days in diverse sequences. From personalized customer information, to improved travel demand models, understanding this type of heterogeneity among transit users is relevant to a number of applications core to public transport agencies’ function. In this study, passenger heterogeneity is investigated based on a longitudinal representation of each user’s multi-week activity sequence derived from smart card data. We propose a methodology leveraging this representation to identify clusters of users with similar activity sequence structure. The methodology is applied to a large sample (n = 33,026) from London’s public transport network, in which each passenger is represented by a continuous 4-week activity sequence. The application reveals 11 clusters, each characterized by a distinct sequence structure. Socio-demographic information available for a small sample of users (n = 1973) is combined to smart card transactions to analyze associations between the identified patterns and demographic attributes including passenger age, occupation, household composition and income, and vehicle ownership. The analysis reveals that significant connections exist between the demographic attributes of users and activity patterns identified exclusively from fare transactions.     

Assessment of spatial transferability of an activity-based model,TASHA

Spatial transferability has been recognized as a useful validation test for travel demand models. To date, however, transferability of activity-based models has not been frequently assessed. This paper assesses the spatial transferability of an activity-based model, TASHA (Travel Activity Scheduler for Household Agents), which has been developed for the Greater Toronto Area (GTA), Canada. TASHA has been transferred to the context of the Island of Montreal, Canada using the 2003 Origin–Destination (O–D) travel survey and the 2001 Canadian Census. It generates daily schedules of activities (individual and joint) for each individual in this region. The modelled activity attributes (frequency, start time, duration and distance) from TASHA and observed attributes from the 2003 O–D travel survey are compared for five different activities (i.e. work, school, shopping, other, and return to home). At the aggregate level, TASHA provides quite reasonable outcomes (in some cases – better results than for the Toronto Area) for all four attributes for work, school and return to home activities with few exceptions (for instance, school start time). The model outcomes are also promising for shopping frequency and start times; however, TASHA provides larger differences for average shopping durations and distances. Only the forecasts for all four attributes for the ‘other’ activity type differ greatly with the observed attributes for the Montreal Island. These large differences most likely indicate the differences in behaviour between the Montreal Island and the Toronto Area. In general, we conclude that re-estimation of model parameters and the use of local activity attribute distributions (frequency, start time and duration) is a desirable step in the transfer of the TASHA model from one context to another.     

Estimation of a constrained multinomial logit model

Identifying the set of available alternatives in a choice process after considering an individual’s bounds or thresholds is a complex process that, in practice, is commonly simplified by assuming exogenous rules in the choice set formation. The Constrained Multinomial Logit (CMNL) model incorporates thresholds in several attributes as a key endogenous process to define the alternatives choice/rejection mechanism. The model allows for the inclusion of multiple constraints and has a closed form. In this paper, we study the estimation of the CMNL model using the maximum likelihood function, develop a methodology to estimate the model overcoming identification problems by an endogenous partition of the sample, and test the model estimation with both synthetic and real data. The CMNL model appears to be suitable for general applications as it presents a significantly better fit than the MNL model under constrained behaviour and replicates the MNL estimates in the unconstrained case. Using mode choice real data, we found significant differences in the values of times and elasticities between compensatory MNL and semi-compensatory CMNL models, which increase as the thresholds on attributes become active.     

Trip chaining propensity and tour mode choice of out-of-home workers: evidence from a mid-sized Canadian city

Suburban development patterns, flexible work hours, and increasing participation in out-of-home activities are making the travel patterns of individuals more complex, and complex trip chaining could be a major barrier to the shift from drive-alone to public transport. This study introduces a cohort-based approach to analyse trip tour behaviors, in order to better understand and model their relationships to socio-demographics, trip attributes, and land use patterns. Specifically, it employs worker population cohorts with homogenous activity patterns to explore differences and similarities in tour frequency, trip chaining, and tour mode choices, all of which are required for travel demand modeling. The paper shows how modeling of these important tour variables may be improved, for integration into an activity-based modeling framework. Using data from the Space–Time Activity Research (STAR) survey for Halifax, Canada, five clusters of workers were identified from their activity travel patterns. These were labeled as extended workers, 8 to 4 workers, shorter work-day workers, 7 to 3 workers, and 9 to 5 workers. The number of home-based tours per day for all clusters were modeled using a Poisson regression model. Trip chaining was then modeled using an Ordered Probit model, and tour mode choice was modeled using a Multinomial logit (MNL) model. Statistical analysis showed that socio-demographic characteristics and tour attributes are significant predictors of travel behavior, consistent with existing literature. Urban form characteristics also have a significant influence on non-workers’ travel behavior and tour complexity. The findings of this study will assist in the future evaluation of transportation projects, and in land-use policymaking.

     

Forecasting travel behavior using Markov Chains-based approaches

Recent advances in agent-based micro-simulation modeling have further highlighted the importance of a thorough full synthetic population procedure for guaranteeing the correct characterization of real-world populations and underlying travel demands. In this regard, we propose an integrated approach including Markov Chain Monte Carlo (MCMC) simulation and profiling-based methods to capture the behavioral complexity and the great heterogeneity of agents of the true population through representative micro-samples. The population synthesis method is capable of building the joint distribution of a given population with its corresponding marginal distributions using either full or partial conditional probabilities or both of them simultaneously. In particular, the estimation of socio-demographic or transport-related variables and the characterization of daily activity-travel patterns are included within the framework. The fully probabilistic structure based on Markov Chains characterizing this framework makes it innovative compared to standard activity-based models. Moreover, data stemming from the 2010 Belgian Household Daily Travel Survey (BELDAM) are used to calibrate the modeling framework. We illustrate that this framework effectively captures the behavioral heterogeneity of travelers. Furthermore, we demonstrate that the proposed framework is adequately adapted to meeting the demand for large-scale micro-simulation scenarios of transportation and urban systems.     

Urban activity pattern classification using topic models from online geo-location data

Location-based check-in services in various social media applications have enabled individuals to share their activity-related choices providing a new source of human activity data. Although geo-location data has the potential to infer multi-day patterns of individual activities, appropriate methodological approaches are needed. This paper presents a technique to analyze large-scale geo-location data from social media to infer individual activity patterns. A data-driven modeling approach, based on topic modeling, is proposed to classify patterns in individual activity choices. The model provides an activity generation mechanism which when combined with the data from traditional surveys is potentially a useful component of an activity-travel simulator. Using the model, aggregate patterns of users’ weekly activities are extracted from the data. The model is extended to also find user-specific activity patterns. We extend the model to account for missing activities (a major limitation of social media data) and demonstrate how information from activity-based diaries can be complemented with longitudinal geo-location information. This work provides foundational tools that can be used when geo-location data is available to predict disaggregate activity patterns.     

Strategic sampling for large choice sets in estimation and application

Many discrete choice contexts in transportation deal with large choice sets, including destination, route, and vehicle choices. Model estimation with large numbers of alternatives remains computationally expensive. In the context of the multinomial logit (MNL) model, limiting the number of alternatives in estimation by simple random sampling (SRS) yields consistent parameter estimates, but estimator efficiency suffers. In the context of more general models, such as the mixed MNL, limiting the number of alternatives via SRS yields biased parameter estimates. In this paper, a new, strategic sampling scheme is introduced, which draws alternatives in proportion to updated choice-probability estimates. Since such probabilities are not known a priori, the first iteration uses SRS among all available alternatives. The sampling scheme is implemented here for a variety of simulated MNL and mixed-MNL data sets, with results suggesting that the new sampling scheme provides substantial efficiency benefits. Thanks to reductions in estimation error, parameter estimates are more accurate, on average. Moreover, in the mixed MNL case, where SRS produces biased estimates (due to violation of the independence of irrelevant alternatives property), the new sampling scheme appears to effectively eliminate such biases. Finally, it appears that only a single iteration of the new strategy (following the initialization step using SRS) is needed to deliver the strategy’s maximum efficiency gains.     

The Impacts of Ict on leisure Activities and Travel: A Conceptual Exploration

This paper offers a conceptual exploration of the potential impacts of ICTs on leisure activities and the associated travel. We start by discussing what leisure is and is not. We point out that the boundaries between leisure, mandatory, and maintenance activities are permeable, for three reasons: the multi-attribute nature of a single activity, the sequential interleaving of activity fragments, and the simultaneous conduct of multiple activities (multitasking). We then discuss four kinds of ways by which ICT can affect leisure activities and travel: the replacement of a traditional activity with an ICT counterpart, the generation of new ICT activities (that may displace other activities), the ICT-enabled reallocation of time to other activities, and ICT as a facilitator of leisure activities. We suggest 13 dimensions of leisure activities that are especially relevant to the issue of ICT impacts: location (in)dependence, mobility-based versus stationary, time (in)dependence, planning horizon, temporal structure and fragmentation, possible multitasking, solitary versus social activity, active versus passive participation, physical versus mental, equipment/media (in)dependence, informal versus formal arrangements required, motivation, and cost. The primary impact of ICT on leisure is to expand an individual’s choice set; however whether or not the new options will be chosen depends on the attributes of the activity (such as the 13 identified dimensions), as well as those of the individual. The potential transportation impacts when the new options are chosen are ambiguous.     

Lifestyle classifications with and without activity-travel patterns

Trip-based approach and activity-based approach are two extremes in the use of activity related information when developing travel demand models. Creating lifestyle clusters for a population is a compromise between the two. On the one hand, it has taken into account travel-activity patterns in the development of the clusters. On the other hand, the clusters represent homogenous groups of individuals and simple activity-based travel demand models can be developed for each cluster. However, the development of such clusters requires knowledge of activity-travel patterns of individuals, which can only be obtained from a large-scale survey. It is still an open question how to create travel/activity-related lifestyle clusters using readily available socio-demographic data (such as census data) alone. This paper attempts to answer this question by proposing a procedure of lifestyle classification that moves from specific surveys to a general population. This paper first studies issues related to the development of homogeneous clusters using socio-economic, demographic and activity-travel data. The second part of the paper addresses the issue of data insufficiency and points out that in order to use the clusters developed for travel demand estimation, it is important to know how to allocate individuals in the population to the developed clusters. As a first attempt, this paper proposes to use a recently developed technique called, Support Vector Machine (SVM), to develop classification functions that based on readily available information only. The methodologies proposed are applied to a sub-urban area in Hong Kong. Six lifestyle clusters are first produced using factor analysis and cluster analysis. SVM is then used to develop classification functions that are based on fewer variables. Results show that the two sets of lifestyle clusters are similar and that the SVM outperforms other traditional classification methods.     

Analysis of consumer response to fuel price fluctuations applying sample selection model to GPS panel data: Dynamics in individuals’ car use

The effects of fuel price increases on people’s car use have been widely discussed during the last few decades in travel behavior research. It is well recognized that fuel price has significant effects on driving distance and driving efficiency. However, most of this research assumed that these effects are invariant across individuals and weather conditions. Moreover, intrinsic variability in people’s preferences has not been given much attention due to the difficulty of collecting the necessary data. In this paper, we collected detailed travel behavior data of 276 respondents in the Netherlands, spanning a time period between one week and three months using GPS logs. These GPS data were fused with weather data, allowing us to estimate both exogenous (such as weather and fuel price) and endogenous effects (inertia and activity plans) on individual’s car use behavior. To further understand the effects of fuel price on the environment, we estimated the effects of fuel price fluctuation on CO₂ emissions by car. The results show a significant degree of inertia in car use behavior in response to increased fuel prices. Weather and fuel price showed significant effects on individual’s car using behavior. Moreover, fuel price shows two-week lagged effects on individual’s travel duration by car.     

Examining the effects of out-of-home and in-home constraints on leisure activity participation in different seasons of the year

Using multi-day, multi-period travel diaries data of 56 days (four waves of two-week diaries) for 67 individuals in Stockholm, this study aims to examine the effects of out-of-home and in-home constraints (e.g. teleworking, studying at home, doing the laundry, cleaning and taking care of other household member[s]) on individuals’ day-to-day leisure activity participation decisions in four different seasons. This study also aims to explore the effects of various types of working schedules (fixed, shift, partial- and full-flexible) on individuals’ decisions to participate in day-to-day leisure activities. A pooled model (56 days) and wave-specific models (14 days in each wave) are estimated by using dynamic ordered Probit models. The effects of various types of working schedules are estimated by using 28 days of two waves’ data. The results show that an individual’s leisure activity participation decision is significantly influenced by out-of-home work durations but not influenced by in-home constraints, regardless of any seasons. Individuals with shift working hours engage less in day-to-day leisure activities than other workers’ types in both spring and summer seasons. The thermal indicator significantly affects individuals’ leisure activity participation decisions during the autumn season. Individuals exhibit routine behaviour characterized by repeated decisions in participating in day-to-day leisure activities that can last up to 14 days, regardless of any seasons.     

Social networks and joint/solo activity–travel behavior

The social dimension of activity–travel behavior has recently received much research attention. This paper aims to make a contribution to this growing literature by investigating individuals’ engagements in joint activities and activity companion choices. Using activity–travel diary data collected in Hong Kong in 2010, this study examines the impact of social network attributes on the decisions between solo and joint activities, and for joint activities, the choices of companions. Chi-square difference tests are used to assess the importance of social network variables in explaining joint activity behavior. We find that the inclusion of social network attributes significantly improves the goodness-of-fit of the model with only socioeconomic variables. Specifically, individuals receiving emotional support and social companionship from family members/relatives are found to more likely undertake joint activities with their family members/relatives; the size of personal social networks is found to be a significant determinant of companion choices for joint activities; and activity companions are found to be significant determinants of travel companions. The findings of this study improve the understanding about activity–travel, especially joint activity–travel decisions.     

Activity-based disaggregate travel demand model system with activity schedules

We present an integrated activity-based discrete choice model system of an individual’s activity and travel schedule, for forecasting urban passenger travel demand. A prototype demonstrates the system concept using a 1991 Boston travel survey and transportation system level of service data. The model system represents a person’s choice of activities and associated travel as an activity pattern overarching a set of tours. A tour is defined as the travel from home to one or more activity locations and back home again. The activity pattern consists of important decisions that provide overall structure for the day’s activities and travel. In the prototype the activity pattern includes (a) the primary – most important – activity of the day, with one alternative being to remain at home for all the day’s activities; (b) the type of tour for the primary activity, including the number, purpose and sequence of activity stops; and (c) the number and purpose of secondary – additional – tours. Tour models include the choice of time of day, destination and mode of travel, and are conditioned by the choice of activity pattern. The choice of activity pattern is influenced by the expected maximum utility derived from the available tour alternatives.     

Collecting a multi-dimensional three-weeks household time-use and activity diary in the Bandung Metropolitan Area,Indonesia

This paper describes a comprehensive panel data collection and analysis at household level, including detailed travel behaviour variables and comprehensive in-home and out-of-home activities, individual cognitive habits and affective behaviours, the rate of physical activity, as well as health related quality of life (QoL) information in the Bandung Metropolitan Area (BMA) of Indonesia. To our knowledge, this is the first attempt to collect an individual’s activity diary over an extended period as it captures the multi-tasking activities and multidisciplinary factors that underlie individual activity-travel patterns in a developing country. Preliminary analyses of the collected data indicate that different beliefs, anticipated emotions, support and attachment to motorised modes significantly correlate with different groups of occupation, gender, age, activity participation, multi-tasking activities, and physical health, but not with different social and mental health. This finding highlights the reason why implementing car reduction policies in Indonesia, without breaking or changing the individual’s habits and influencing his/her attitudes have not been fruitful. The results also show that endorsing more physical activities may result in a significant reduction in the individual’s motorised mode use, whilst individuals who demonstrate a tendency to use their spare time on social activities tend to have better social health conditions. Furthermore, undertaking multi-tasking out-of-home discretionary activities positively correlates with better physical health. All these highlight the importance of properly understanding and analysing the complex mechanisms that underlie these fundamental factors that shape individual daily activity-travel patterns in developing countries. This type of multidisciplinary approach is needed to design better transport policies that will not only promote better transport conditions, but also a healthier society with a better quality of life.