Observing individual dynamic choices of activity chains from location-based crowdsourced data

The existing efforts on studying human mobility and activity using location-based crowdsourced data mainly focus on obtaining the activity chain pattern in a region at an aggregate level. To observe individual dynamic choices of activity chains, this paper presents a data-driven approach to estimating individual-specific activity chain set and corresponding choice probabilities for a given person over a 24-h period using crowdsourced data from location-based service apps. We detect an individual-specific stochastic activity set using a contextual-parcel data analysis. Based on the time geography theory, we refine a space-time bicone concept to construct an activity-travel space-time-state network from the stochastic activity set. These space-time bicone constraints define a set of potential activity choices to reduce the search space of activity location and duration choices. We construct an activity state transition graph from the space-time-state network and calculate a Markov matrix for activity choice probabilities. Furthermore, we calculate the probabilities of activity chain choices using the Markov matrix. We also visualize individual-specific activity chain set in a space-time-state network to show the dynamic choices of individual daily mobility and activity. We demonstrate the proposed approach through conducting numerical analyses using crowdsourced data from location-based service apps - Foursquare and Twitter to construct individual-specific activity choice sets and corresponding choice probabilities.     

Learning daily activity patterns with probabilistic grammars

Daily activity pattern is the reflection and abstraction of actual individual activity participation on daily basis. It carries information on activity type, frequency and sequence. Preference of daily activity patterns varies among population, and thus can be interpreted as personal life styles. This paper advances studies on human daily activity patterns by providing new perspective and methodology in the modeling and learning of daily activity patterns using probabilistic context-free grammars. In this paper, similarities between daily activity pattern—which is defined as activity sequence—and language are explored. We developed context-free grammars to parse and generate daily activity patterns. To replicate people’s heterogeneity in selecting daily activity patterns, we introduced probabilistic context-free grammars and proposed several formulations to estimate the probability of a context-free grammar with daily activity patterns observed in household travel survey. We conducted experiments on the proposed formulations, finding that under proper context-free grammar and problem formulation, the estimated probabilistic context-free grammar is able to reproduce the observed pattern distribution in household travel survey with satisfactory precision. Practically, the proposed methodology sheds light on the issue of generating stochastic and accessibility-dependent choice sets for daily activity pattern models in certain activity-based modeling frameworks.     

Activity planning processes in the Agent-based Dynamic Activity Planning and Travel Scheduling (ADAPTS) model

This paper describes the representation of the activity planning process utilized in a new activity-based microsimulation model called the ADAPTS (Agent-based Dynamic Activity Planning and Travel Scheduling) model, which dynamically simulates activity and travel planning and scheduling. The model utilizes a dynamic activity planning framework within the larger overall microsimulation system, which is a computational process model that attempts to replicate the decisions which comprise time-dependent activity scheduling. The model presents a step forward in which the usual concepts of activity generation and activity scheduling are significantly enhanced by adding an additional component referred to as activity planning in which the various attributes which describe the activity are determined. The model framework, therefore, separates activity planning from activity generation and treats all three components, generation, planning and scheduling, as separate discrete but dynamic events within the overall microsimulation. The development of the planning order model, which determines when and in what order each activity planning decision is made is the specific focus of this paper. The models comprising the planning order framework are developed using recent survey data from a GPS-based prompted recall survey. The model development, estimation, validation, and its use within the overall ADAPTS system are discussed. A significant finding of the study is the verification of the apparent transferability of the activity planning order model.     

Repetitions in individual daily activity–travel–location patterns: a study using the Herfindahl–Hirschman Index

Using Herfindahl–Hirschman Index and the Mobidrive and Thurgau six-week travel diary datasets this paper examines the degree of repetition of individuals’ choices of their daily activity–travel–location combinations. The results show that the repetitiveness of individual activity–travel–mode–location combinations is highly influenced by the individuals’ out-of-home commitments, the intra-household conditions and the availability and the accessibility of the activity locations. Different types of activity have different pattern of repetition. The level of repetition of individual’s daily activity–travel pattern is less correlated to travel mode choice, but more to the individuals’ commitments and obligations. The repetitiveness of mode choices is more related to the conditions or the accessibilities of the activity location, but not directly to the activity itself.     

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.     

Well-being and activity-based models

We present empirical and theoretical analyses to investigate the relationship between happiness (or subjective well-being) and activity participation and develop a framework for using well-being data to enhance activity-based travel demand models. The overriding hypothesis is that activities are planned and undertaken to satisfy needs so as to maintain or enhance subjective well-being. The empirical analysis consists of the development of a structural equations exploratory model of activity participation and happiness using data from a cross-sectional survey of a sample of commuters. The model reveals significant correlations between happiness and behavior for different types of activities: higher propensity of activity participation is associated with greater activity happiness and greater satisfaction with travel to the activity. The theoretical analysis consists of the development of a modeling framework and measures for the incorporation of well-being within activity-based travel demand models. The motivation is that activity pattern models have been specified in ad-hoc ways in practice as a function of mobility, lifestyle, and accessibility variables. We postulate that well-being is the ultimate goal of activity patterns which are driven by needs and propose two extensions of activity pattern models. The first extension consists of the use of well-being measures as indicators of the utility of activity patterns (in addition to the usual choice indicators) within a random utility modeling framework. The second extension models conceptually the behavioral process of activity generation based on needs satisfaction. We present an example of an operational activity pattern model and propose well-being measures for enhancing it.     

Activity space estimation with longitudinal observations of social media data

In this paper, we demonstrate the use of an inexpensive and easy-to-collect long-term dataset to address the problems caused by basing activity space studies off short-term data. In total, we use 63,114 geo-tagged tweets from 116 unique users to create individuals’ activity spaces based on minimum bounding geometry (convex hull). By using polygon density maps of activity space, we found clear differences between weekday and weekend activity spaces, and were able to observe the growth trajectory of activity space over 17 weeks. In order to reflect the heterogeneous nature of spatial behavior and tweeting habits, we used Latent Class Analysis twice. First, to identify five unique patterns of location-based activity spaces that are different in shape and anchoring. Second, we identify three unique growth trajectories. The comparison among these latent growth trajectories shows that in order to capture the extent of activity spaces we need long time periods for some individuals and shorter periods of observation for others. We also show that past studies using a single digit number of weeks may not be sufficient to capture individuals’ activity space. The major activity locations identified using a multilevel latent class model, do not appear to be statistically related to the growth patterns of Twitter users activity spaces. The evidence here shows Twitter data can be a valuable complementary source of information for heterogeneity analysis in activity-based modeling and simulation.     

Modeling activity scheduling time horizon: Duration of time between planning and execution of pre-planned activities

Growing recognition that observed travel patterns are the result of an underlying activity scheduling process has resulted in a new stream of data collection and modeling efforts. Of particular focus is the planning or sequencing of activity scheduling decisions over time that precede actual execution of activities/trips. Understanding and potentially modeling these sequences offers particular promise, as strong interdependencies in activity/travel choices likely exist. In practice, however, a fixed order of sequencing by activity type is often assumed that overlooks the strong interdependencies in activity/travel choices and can be misleading. This study presents the process of developing parametric and non-parametric hazard models to predict the duration of time between planning and execution of pre-planned activities based on attributes of activity and characteristics of decision maker. Modeling results suggest that activity type alone may not suffice to fully explain how activities are planned. Rather, the nature of the activity and several overriding personal and situational factors play an important role. This will make the model more amenable to a variety of people and situations and will make it more sensitive to emerging policy action scenarios.     

A structural equations analysis of commuters' activity and travel patterns

An exploratory analysis of commuters' activity and travel patterns was carried out using activity-based travel survey data collected in the Washington, DC metropolitan area to investigate and estimate relationships among socio-demographics, activity participation, and travel behavior. Structural equations modeling methodology was adopted to determine the structural relationships among commuters' demographics, activity patterns, trip generation, and trip chaining information. Three types of structural equations model systems were estimated: one that models relationships between travel and activity participation, another that captures trade-offs between in-home and out-of-home activity durations, and a third that models the generation of complex work trip chains. The model estimation results show that strong relationships do exist among commuters' socio-demographic characteristics, activity engagement information, and travel behavior. The finding that significant trade-offs exist between in-home and out-of-home activity participation is noteworthy in the context of in-home vs. out-of-home substitution effects. Virtually all of the results obtained in this paper corroborate earlier findings reported in the literature regarding relationships among time use, activity participation, and travel. This revised version was published online in June 2006 with corrections to the Cover Date.     

A random utility maximization (RUM) based dynamic activity scheduling model: Application in weekend activity scheduling

The paper presents a modeling framework for dynamic activity scheduling. The modeling framework considers random utility maximization (RUM) assumption for its components in order to capture the joint activity type, location and continuous time expenditure choice tradeoffs over the course of the day. The dynamics of activity scheduling process are modeled by considering the history of activity participation as well as changes in time budget availability over the day. For empirical application, the model is estimated for weekend activity scheduling using a dataset (CHASE) collected in Toronto in 2002–2003. The data set classifies activities into nine general categories. For the empirical model of a 24-h weekend activity scheduling, only activity type and time expenditure choices are considered. The estimated empirical model captures many behavioral details and gives a high degree of fit to the observed weekend scheduling patterns. Some examples of such behavioral details are the effects of time of the day on activity type choice for scheduling and on the corresponding time expenditure; the effects of travel time requirements on activity type choice for scheduling and on the corresponding time expenditure, etc. Among many other findings, the empirical model reveals that on the weekend the utility of scheduling Recreational activities for later in the day and over a longer duration of time is high. It also reveals that on the weekend, Social activity scheduling is not affected by travel time requirements, but longer travel time requirements typically lead to longer-duration social activities.     

Real-time tracking of activity scheduling/schedule execution within a unified data collection framework

One of the major foci in transport research is the identification of the temporal–spatial decision making structure embedded in activity scheduling and its linkage to actual activity execution. The latter part of the research in question has not been explored explicitly in real life situations due to the lack of effective data collection means. This paper presents a real-time activity scheduling, activity/travel survey system that incorporates the extraction of activity scheduling and the execution information within one unified data collection framework. These “revealed” data can be used for explicitly defining the mechanism of how people’s activity schedules dynamically adapt to social-demographic and temporal–spatial constraints and finally lead to the observed activity-travel patterns.     

Trend analysis of activity generation attributes over time

Activity generation models are relatively poorly developed in activity-based travel demand modelling frameworks. This research investigates whether observed distributions of activity attributes (activity frequency, start time and duration) used as inputs in the activity generation component of an activity-based travel demand model have changed over time. This research empirically examines changes in the distributions of activity generation attributes over time in the Greater Montreal area (GMA), Quebec, Canada. It also focuses on how these attributes vary with peoples’ socio-demographic characteristics. This research relies on the 1998, 2003 and 2008 origin–destination (O–D) household travel surveys of the GMA. The comparative analysis at three time points in a 10-year period clearly reveals that distributions of activity attributes are significantly changing over time. Work and school activities show similar trends; frequency “1” has increased and frequency “2+” has decreased over time. The occurrence of shopping activity on weekdays is decreasing over time. Start time and duration distributions for each activity have also changed significantly over time. The research allows preparing activity attributes for the application of an activity-based model, TASHA, such that they reflect temporal changes in travel behaviour of the GMA.     

Daily activity pattern recognition by using support vector machines with multiple classes

The focus of this paper is to learn the daily activity engagement patterns of travelers using Support Vector Machines (SVMs), a modeling approach that is widely used in Artificial intelligence and Machine Learning. It is postulated that an individual’s choice of activities depends not only on socio-demographic characteristics but also on previous activities of individual on the same day. In the paper, Markov Chain models are used to study the sequential choice of activities. The dependencies among activity type, activity sequence and socio-demographic data are captured by employing hidden Markov models. In order to learn model parameters, we use sequential multinomial logit models (MNL) and multiclass Support Vector Machines (K-SVM) with two different dependency structures. In the first dependency structure, it is assumed that type of activity at time ‘t’ depends on the last previous activity and socio-demographic data, whereas in the second structure we assume that activity selection at time ‘t’ depends on all of the individual’s previous activity types on the same day and socio-demographic characteristics. The models are applied to data drawn from a set of California households and a comparison of the accuracy of estimation of activity types and their sequence in the agenda, indicates the superiority of K-SVM models over MNL. Additionally, we show that accuracy in estimating activity patterns increases using different sets of explanatory variables or tuning parameters of the kernel function in K-SVM.     

Redefining activity types: Who participates in which leisure activity?

This paper analyzes the activity choices of individuals and the links between socio-demographics, daily schedules and activity attributes using a new activity choice framework. Activities are first clustered into groups based on their salient attributes, such as duration, frequency, flexibility, planning times, and number of involved persons, rather than their functional types (work, leisure and household obligations), using a K-means cluster technique. This led to the creation of several new activity groups such as “long, temporally fixed, personally flexible activities”, “short and flexible activities”. These activity groups form the choice set for the mixed logit activity choice modeling structure developed for the leisure activities in the second part of the paper. The model results reveal the significant relationships between socio-demographics, temporal characteristics, and characteristics of the schedules on leisure activity choice. The results demonstrate how changing demographics and other activities in individuals’ schedules may affect the nature of the leisure activities and present the substitution and complimentary effects that these new activity groups have on one another.     

A conceptual framework of individual activity program generation

The research in this paper attempts to better understand the process by which activities are generated at an individual level. Activity-based travel analyses have gained popularity in recent years because they recognize the complexity of activity behavior and view travel as a derivative of this behavior. Most activity-based studies have focused on the spatial and temporal linkage of trips; that is, the scheduling of activities. They consider the agenda of activities for participation, and associated attributes of the activity participation (such as mode to activity and location of activity performance), as predetermined. This paper develops a comprehensive conceptual framework of the relatively unexplored area of activity agenda generation. Such a framework will be valuable in empirical modeling of activity generation behavior. A subsequent paper focuses on translating a part of this conceptual framework into an empirical model.     

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.     

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.

A combined activity/travel choice model for congested road networks with queues

This paper presents a combined activity/travel choice model and proposes a flow-swapping method for obtaining the model's dynamic user equilibrium solution on congested road network with queues. The activities of individuals are characterized by given temporal utility profiles. Three typical activities, which can be observed in morning peak period, namely at-home activity, non-work activity on the way from home to workplace and work-purpose activity, will be considered in the model. The former two activities always occur together with the third obligatory activity. These three activities constitute typical activity/travel patterns in time-space dimension. At the equilibrium, each combined activity/travel pattern, in terms of chosen location/route/departure time, should have identical generalized disutility (or utility) experienced actually. This equilibrium can be expressed as a discrete-time, finite-dimensional variational inequality formulation and then converted to an equivalent "zero-extreme value" minimization problem. An algorithm, which iteratively adjusts the non-work activity location, corresponding route and departure time choices to reach an extreme point of the minimization problem, is proposed. A numerical example with a capacity constrained network is used to illustrate the performance of the proposed model and solution algorithm.     

Activity–travel behaviour of non-workers from Bangalore City in India

This paper presents exploratory and statistical analyses of the activity–travel behaviour of non-workers in Bangalore city in India. The study summarises the socio-demographic characteristics as well as the activity–travel behaviour of non-workers using a primary activity–travel survey data collected by the authors. Where possible, the research also compares the analysis findings with the case studies on activity–travel behaviour of non-workers, carried out in developed and developing countries. This gives an opportunity to understand the differences/similarities in the activity–travel behaviour of non-workers across diverse socio-cultural settings. The preliminary exploratory analysis shed light on the differences in activity participation, trip chaining, time-of-day preference for trip departure, and mode use behaviour of non-workers in Bangalore city. Statistical models were developed for investigating the effects of individual and household socio-demographics, land use parameters, and travel context attributes on activity participation, trip chaining, time-of-day choice, and mode choice decisions of non-workers. A few important results of the analysis are the influence of viewing television at home on out-of-home activity participation and trip-chaining behaviour, and the impact of in-home maintenance activity duration on time-of-day choice. Further, based on the findings of the initial analyses, an attempt has been made in this study to develop an integrated model that links time allocation, time-of-day choice, and trip chaining behaviour of non-workers. The study also discusses the implications of the research findings for transportation planning and policy for Bangalore city.     

Modeling household interactions in daily in-home and out-of-home maintenance activity participation

The activity travel patterns of individuals in a household are inter-related, and the realistic modeling of activity-travel behavior requires that these interdependencies be explicitly accommodated. This paper examines household interactions impacting weekday in-home and out-of-home maintenance activity generation in active, nuclear family, households. The in-home maintenance activity generation is modeled by examining the duration invested by the male and female household heads in household chores using a seemingly unrelated regression modeling system. The out-of-home maintenance activity generation is modeled in terms of the decision of the household to undertake shopping, allocation of the task to one or both household heads, and the duration of shopping for the person(s) allocated the responsibility. A joint mixed-logit hazard-duration model structure is developed and applied to the modeling of out-of-home maintenance activity generation. The results indicate that traditional gender roles continue to exist and, in particular, non-working women are more likely to share a large burden of the household maintenance tasks. The model for out-of-home maintenance activity generation indicates that joint activity participation in the case of shopping is motivated by resource (automobiles) constraints. Finally, women who have a higher propensity to shop are also found to be inherently more efficient shoppers.