Examining virtual–physical activity links across purposes, timeframes and locations in the SuperApps era

Abstract

This study examines time allocation and associations between virtual and physical activities across purposes, times, and locations, considering users’ socio-demographic characteristics, residential locations, and personality traits, while also accounting for multitasking. It uses one-week time-use diaries from smartphone users in four Indonesian cities, analyzed with multivariate ordered regression models. The results reveal strong associations between virtual and physical activities across mandatory, maintenance, and leisure purposes. Leisure activities appear less flexible than other types. Multitasking is positively associated with primary activities, especially maintenance tasks like e-shopping, while work-related virtual tasks show negative associations with other activities, reflecting remote work flexibility. Virtual activity engagement on weekdays is negatively linked to physical activity but weakens on weekends. Urban residents, and younger, wealthier individuals are more engaged in virtual activities, whereas males and those in larger households prioritise physical activities. Creative individuals prefer physical leisure, while sociable individuals balance virtual and physical activities.

Introduction

An understanding of how individuals strategically arrange their activities across time and space is fundamental in order to formulate suitable and effective urban policies in various domains such as transportation, health, and energy. In this context, time use has been a subject of interest to many researchers for decades. In particular, previous studies have investigated the variability in time use across different types of activities, days, locations, and individuals’ socio-demographic characteristics^{1, 2–3}. The long-standing paradigm of time and space constraints⁴ has provided a framework for evaluating the ways in which different opportunities and constraints shape individuals’ participation in activities across time and space, as well as underlining the variability arising from intra- and interpersonal differences between individuals.

The rise in the use of information and communication technologies (ICTs) has relaxed these constraints, however, allowing individuals to perform virtual activities that have a different nature from physical activities. ICTs offer various opportunities, such as participating in or rearranging activities without needing to be physically present, meaning that an individual’s needs can be met at anytime and anywhere, free from traditional physical and authority-related constraints⁵. Consequently, individuals now continuously adjust their daily routines through trade-offs between online and offline engagements. These evolving behaviour patterns have broad implications in terms of travel demand and the landscapes of urban infrastructure. As virtual activities become more prevalent, there is an increasingly urgent need to investigate the factors that are driving these changes. This knowledge could help policymakers develop strategies to effectively manage their societal and environmental impacts.

Prior studies have investigated the impact of ICTs on activity participation and most have used a framework that considers four impacts from ICT on travel and participation in activities⁶. Substitution effects involve avoiding, replacing, or eliminating travel through the use of ICT services, whereas generation effects increase travel demand via ICTs. Neutral implication means that ICTs have no travel impact, while modification means that the use of ICTs alters travel behaviour, such as changes in mode or timing. These represent the various ways in which individuals rearrange their activities and have been used to investigate the impact of various ICT services. E-shopping and teleworking have been the main subjects of interest in prior studies and can affect activity participation in various ways. For instance, prior studies have found that e-shopping can be both a substitute for^7,8 and complementary to⁹ needs for physical shopping activity. Teleworking has also been found to both replace work trips¹⁰ and to generate additional trips¹¹. However, most of these prior studies have focused on the trade-off between the virtual and physical aspects of a specific activity, such as e-shopping and in-store shopping.

In many circumstances, there are relationships between engagement in virtual activities and broader activity participation (due to the rebound effects of ICTs). These arise because the travel time and cost savings provided by the use of ICT services can be allocated to new activities in various purposes. For instance, e-shopping provides opportunities to perform other activities, due to savings in travel time¹², while teleworking allows individuals to rearrange their household or leisure activities¹³. Various studies have investigated the rebound effects of ICTs^14,15 and have found various mechanisms for the rearrangement of activities. Nevertheless, many studies in this field have only examined the effects of specific ICT platforms, with limited focus on integrated spatiotemporal variables. Individuals now use various ICT platforms to organise their daily schedules across different locations and times, and the relationships between virtual and physical activities are complex.

Furthermore, the intensity of virtual activities has grown substantially. In 2023, there were 5.35 billion internet users, making up about 66% of the world’s population, and on average, each user spends 6.5 h a day online¹⁶. This usage is driven by the increased availability and advancement of ICTs. The smartphone is one of the most popular forms of ICT. Compared to apps that are available for wired ICTs, such as personal computers and fax machines, smartphone apps are more versatile and flexible and offer various functions that can be used at anytime and anywhere. They also facilitate multitasking by allowing multiple activities to be conducted simultaneously¹⁷. This flexibility often leads to a more varied distribution of virtual activities across different purposes, places and days.

Despite these conditions, most prior studies examining the interaction between virtual and physical activities have focused on only one dimension at a time, as such investigating virtual activities across different purposes, places, or days separately. For instance, Liu et al.¹⁸ and Bin et al.¹⁹ explored the trade-off between activities based on the purposes of the activity, while Susilo and Liu²⁰ studied time-use allocations while also considering temporal variables, thereby allowing for an examination of mobility patterns between days. More importantly, previous studies have often overlooked the issue of multitasking and have assumed that virtual activities were only conducted as primary/main activities. In practice, however, multitasking has become increasingly common and serves as a key strategy for maximising daily activities. By incorporating multitasking along with additional dimensions such as time, purpose, and location, while also leveraging more comprehensive longitudinal data, the relationship between virtual and physical activities could be better understood.

The main objective of this paper is to explore daily time allocations, their determinants, and the varying relationships between virtual and physical activities. The main contributions of this study are to address an existing research gap by examining the relationship between virtual and physical activities across different purposes, times, and locations, while also incorporating multitasking activities into the analysis. The one-week time-use diaries from smartphone users in four different Indonesian cities is used as data set. Key determinants include users’ personal characteristics, residential locations and personality traits. The data collection process spanned four cities in Indonesia, representing medium-sized, large, metropolitan and megapolitan cities, to ensure a comprehensive understanding of app-use behaviour across different urban settings.

The remainder of this paper is organised as follows. The next section reviews the literature on time allocation and the influence of ICT on time use. Section 3 describes the methods, including the one-week time-use diary that serves as the research dataset, as well as the data analysis methodology. Section 4 presents the data used in the study and the results of the model estimation, focusing on time allocation and the interplay between virtual and physical activities. The paper concludes with a discussion of the findings and suggestions for future research.

Literature Review

Time allocation and the factors influencing It

Individuals engage in a range of activities across time and space to meet their needs and desires, and their strategic decisions are often reflected in how they allocate their daily time use^{4,20, 21–22}. An understanding of time allocation is critical for gaining insights into individual behaviour, and this has been extensively studied across multiple disciplines in order to identify the factors that shape these decisions. This strategic focus on time allocation provides a valuable lens for examining behavioural patterns, thereby informing targeted research and policy development.

Time allocation patterns have been shown to vary based on both individual and contextual factors. Much of the existing research has explored how people distribute their time across different types of activities (for example, mandatory, leisure, and maintenance activities). Mandatory activities tend to remain stable throughout the week, in contrast to leisure and maintenance activities^23,24. This stability is attributed to the structured nature of mandatory activities, which are often governed by formal obligations, such as employment contracts or educational enrolments. Since activities are closely linked to spatial locations, time allocation also varies by place. Several studies have shown that individuals spend more of their time at home compared to other locations such as workplaces^25,26. This is largely because the home serves multiple essential functions (e.g., rest, personal care, social interaction with family, and remote work). Moreover, time allocation patterns often differ across the days of the week. Prior research has found that weekday time use tends to be relatively consistent, while weekend patterns vary more significantly and lean towards more discretionary activities^24,27. Other studies^28,29 have identified relationships between the time allocations for weekdays and weekends, suggesting that behaviours on one type of day may influence or reflect patterns on the other.

Time allocation also has also been examined by investigating the factors that influence it. Researchers have found that daily time allocations are associated with socio-demographic characteristics; for example, younger and wealthier individuals tend to spend more time on out-of-home social and recreational activities^1,30, while men are more likely to engage in outdoor physical leisure activities such as sports or to dedicate more time to work^31,32. Daily time allocation also differs based on household characteristics^33,34. Households with two working members are more likely to allocate their time to out-of-home work-related activities³³, whereas couples with children are less inclined to spend time on out-of-home leisure activities³⁵. Moreover, the urban context plays a role in shaping time allocation, which is influenced by differing economic and cultural settings. A study of eight European cities found that compared to Brussels, residents in French and Swiss cities spend more time working, with the greatest impact observed in Bern and Zurich, followed by Geneva and then the French cities³².

Time allocation in the digital age

Recent studies of time allocation have begun to investigate the influence of ICT use on activity participation. The reason for this is that ICTs enable individuals to participate in activities without being physically present, and to perform multiple tasks simultaneously⁵. These capabilities have reshaped how people allocate their time, as they offer the flexibility to engage in activities either in person or virtually, across diverse locations and schedules, facilitated by ICT. Couclelis³⁶ described this phenomenon as the fragmentation of activities, where tasks that were once confined to specific time blocks and locations are now divided into smaller segments, dispersed across various locations, and often intertwined with other tasks. With the development of ICTs, the duration of virtual daily activities has become substantial, and these now encompass mandatory, leisure, and maintenance activities³⁷.

Various studies have been conducted to understand the relationship between ICTs, time use and travel. The interaction between ICTs and time use is evident across a range of daily activities, including work, shopping, leisure, and social interactions. For instance, increased use of e-shopping services has been found to reduce time spent on out-of-home shopping, particularly for non-daily in-store purchases³⁸. However, the time or cost savings gained in this way have broader implications, as individuals can reallocate these savings to undertake a broad array of tasks. This leads to a multifaceted effect on the expansion of ranges of activity ³⁹. Work by Wang and Law⁴⁰ found that ICT usage can lead to additional time being spent on out-of-home recreational activities. More specifically, e-shopping has been found to create opportunities for engaging in other activities, such as leisure, due to savings in travel time^12,41. Teleworking also influences the time allocated to household tasks and leisure activities¹³. These findings highlight that the effects of ICT usage extend beyond the specific activities it facilitates, suggesting a complex mechanism for the reorganisation of daily activities driven by ICT usage.

Moreover, prior research has investigated the factors influencing ICT usage before investigating the implications on time allocation. From socio-demographic characteristics, it has been found that younger, more educated and affluent households, and those that reside in high-end areas are associated with greater ICT usage, particularly for teleworking, e-shopping, and ride-hailing services^{42, 43, 44, 45, 46–47}. These variables interact with ICT usage to influence time allocation patterns^13,40,48. For instance, a study by Losa Rovira et al.⁴⁹ found that people who are full-time workers and highly educated tend to allocate more time to virtual working activities. Rizki et al.³⁷ reported that people who mostly spend time in their homes have a higher rate of virtual activity. In addition, from attitudinal factors, personality traits have been found to influence individuals’ behaviour in regard to ICT usage^{50, 51–52}; for instance, individuals who are disorganised and impulsive often have a higher rate of smartphone usage. Despite this finding, however, the relationship between these traits and daily time allocation remains poorly understood.

Although many studies have examined the relationships between virtual and physical activities, relatively few have focused on how individuals adjust the locations and timings of their activities in response to the use of virtual platforms. Virtual activities tend to be more widely dispersed, due to more relaxed time-space constraints⁵³, often leading to changes in both the timing and location of physical activities. Despite previous research that has examined the impact of ICTs on time use and activity participation^15,41,54, studies incorporating spatiotemporal variables remain limited^19,20. The few studies that do exist suggest a trade-off between time of day and location, as individuals adjust their schedules in response to the use of virtual platforms for teleworking and e-shopping, for instance. However, most of this research focuses narrowly on certain virtual activities and overlooks the broader range of ICT-enabled virtual engagements. It also assumes that individuals engage in one virtual activity at a time and disregards the potential for multitasking. This study seeks to address these gaps by examining a broader range of virtual activities and integrating multitasking within spatial, contextual, and temporal analyses. By considering these factors, it aims to provide a more comprehensive understanding of how ICT influences daily time allocation and activity patterns.

Methods

Study location

The study was carried out in the Indonesian cities of Jakarta, Bandung, Denpasar, and Cianjur (Fig. 1). These locations were selected because each has unique characteristics in terms of size, population, infrastructure quality, and economic activities, thereby allowing for an analysis of varying physical and virtual activity behaviours^25,55,56. Jakarta is Indonesia’s largest megapolitan area, with a population of over 10 million, and serves as the nation’s economic hub. It features extensive public transport infrastructure, including the Jakarta Mass Rapid Transit and Transjakarta Bus Rapid Transit. Bandung, a metropolitan city known for its education and small industries, has a population of over 2.5 million, and offers bus transit services and the East-West Commuter Train, although these run less frequently than in Jakarta. Denpasar is the capital of Bali, and is renowned for tourism, with a population of over 900,000. It is part of the Greater Denpasar agglomeration and provides bus transit services at a lower frequency than Jakarta. Cianjur, with a population of 160,000, has the smallest area of the cities studied here (26.15 km²), and is primarily industrial and agricultural. It has the fewest public transport services, mainly consisting of paratransit services. While there has been extensive research on travel behaviour in larger cities such as Jakarta, Bandung, and Denpasar, there has been limited research on smaller areas such as Cianjur. This study also aims to provide insights into activity behaviour in smaller cities and how this might affect participation in virtual and physical activities and trade-offs between activities.

Fig. 1 Map of Study Location. [Images not available. See PDF.]

Figure 1 shows the four Indonesian cities included in the study: Jakarta, Bandung, Denpasar, and Cianjur. Jakarta is a megapolitan area with extensive public transport infrastructure and a population of over 10 million, representing Indonesia’s economic center. Bandung, with a population over 2.5 million, is a metropolitan city known for education and small industries. Denpasar, the capital of Bali, has over 900,000 residents and is known for tourism. Cianjur, the smallest city in this study, has approximately 160,000 people and is primarily industrial and agricultural. The map illustrates each city’s geographic location, allowing for comparison of activity patterns and app usage behaviour in different urban contexts.

Instrument

For this study, data were drawn from a cross-dimensional one-week virtual and physical activity diary survey conducted in Jakarta, Bandung, Denpasar and Cianjur in 2022³⁷. The aim of the survey was to gather information from smartphone users about their daily virtual and physical activities and the impact of mobile app usage on their travel behaviours and multitasking habits. Two self-reported questionnaires were used: the first was a one-week time-use and app-use diary, while the second gathered general information on each respondent’s socio-demographic characteristics, personality traits, usage patterns, motivations, attachment to apps and perceptions of their residential environment. A comprehensive description of the full questionnaire is available in previous work³⁷. Data for the study were drawn from the general questionnaire (covering socio-demographic factors, home locations and personality traits) and the time-use diary. Identification of respondents’ personality traits was done based on the Big Five Inventory (BFI) model⁵⁷, which includes questions about five personality traits (extraversion, agreeableness, conscientiousness, neuroticism and openness) measured on a self-reported five-level Likert scale (1: strongly disagree to 5: strongly agree). For each of these five dimensions, there are 19 questions on personality traits, as given in Table 1. The following paragraphs will describe the sections of the time-use diary used in this study.

Table 1. List of personality trait questions

Personality traits. I am a person who…		Measurement
Extraversion	tends to be quiet.	1. Strongly Disagree to 5. Strongly Agree
	is outgoing or sociable.
	is dominant or acts as a leader.
	is hoping other to be a leader
	is full of energy.
Agreeableness	is compassionate or has a soft heart.
	is respectful or treats others with respect.
	assumes the best about people.
Conscientiousness	keeps things neat and tidy.
	tends to be disorganised
	is persistent or works until the task is finished.
	has difficulty getting started on tasks.
	can be somewhat careless.
Negative emotionality	worries a lot.
	tends to feel depressed or blue.
	is temperamental or gets emotional easily.
Open-Mindedness	is fascinated by art, music, or literature
	is complex or a deep thinker.
	is original or comes up with new ideas.

This table outlines the specific statements used to assess respondents' Big Five personality traits (Extraversion, Agreeableness, Conscientiousness, Negative Emotionality, and Open-Mindedness). Each trait includes 3–5 statements, rated by participants on a 5-point Likert scale. These items form the basis for personality analysis in relation to time-use behaviour.

The time-use diary included questions about types of activity, start and end times, locations, and travel details such as time, mode, and cost. Responses were guided by a list of 23 typical daily activities (e.g., eating, shopping, working), and respondents could also write about specific activities that were not listed. These activities were categorised into three groups, mandatory, maintenance, and leisure activities, as defined by Dhamowijoyo et al.²⁵:

• Mandatory activities are those that meet basic human needs, such as sleeping and eating, as well as daily obligations that require interaction with others outside the home. Examples include working in an office, studying at school, or attending a business meeting at a café.

• Maintenance activities involve tasks that are necessary for personal and household upkeep, including chores such as housekeeping, shopping, banking, or receiving medical care.

• Leisure activities cater to social, cultural, and psychological well-being, and encompass entertainment (e.g., watching TV, listening to music), social interactions (e.g. chatting with family or friends, playing sports), vacations, and attending cultural events.

The diary also featured a multitasking report, which tracked individuals’ engagement in multiple activities simultaneously. In this context, multitasking was defined by the respondents themselves as the performance of a secondary activity that occurred concurrently with a main (primary) activity¹⁷. It is important to note that the survey did not measure the level of attention or cognitive load associated with these main/secondary activities. Respondents were asked to describe their primary activities, including the type, location, timing, and cost, and to list any secondary activities occurring at the same time, giving similar details. Secondary activities were always associated with a main activity, and a pair of such activities constituted an occurrence of multitasking. Respondents were also asked to specify whether each activity was virtual (mediated by ICT, such as watching TV or participating in online meetings) or physical (not involving ICT, such as exercising or sleeping). These multitasking categories (secondary and main activities) and activity engagement types (virtual and physical) were used to classify the activities in the data analysis. Further information on the survey design can be found in Rizki et al.³⁷.

Data collection

For each of the four Indonesian cities considered in the study, a dedicated team of interviewers was created. Local coordinators were responsible for recruiting, reviewing and selecting these interviewers. The selected interviewers distributed the questionnaires, with teams assigned to cover each district to ensure comprehensive coverage³⁷. A preliminary evaluation of the questionnaire was done in April 2022, and the final survey was conducted in Bandung, Denpasar, Cianjur and Jakarta, starting in mid-May 2022. Based on feedback from the preliminary survey, both paper-and-pencil and online questionnaires were used. The online questionnaire was used to collect general information (e.g., gender, income, etc.), while the paper-and-pencil approach was used for the app-use diaries, due to the complexity of the questionnaire and the limited quality of ICTs in certain cities. Convenience sampling was employed to gather data. Each city had 30 to 35 interviewers, and each interviewer managed about 5 to 10 respondents. Since the diaries took a week to complete, multiple meetings with respondents were necessary.

The sample size was estimated to reflect the population size of each district, and ethical guidelines were emphasised, including data anonymisation and use of data solely for research purposes. Participants received financial incentives corresponding to local wage standards. Eligibility was limited to individuals aged 18 or older with experience of using apps and the internet. The survey, which was completed by mid-January 2023, initially involved 3092 individuals. To maintain validity, respondents participated in multiple sessions of face-to-face meetings to complete two self-reported questionnaires, following the approach used in previous time-use studies²⁵. These multiple face-to-face meetings were conducted to provide assistance to respondents in filling out the complex time-use diary correctly and to stimulate the remembering process through discussion. They also involved follow-up on the results of the data verification process, including checking for completeness and identifying outliers in terms of activity count, duration, and location. After cleaning and validation, the data from 1193 respondents are used for the analysis.

Valid responses were distributed across the cities as follows: 368 from Jakarta (31%), 336 from Bandung (28%), 269 from Denpasar (23%), and 220 from Cianjur (18%), with the relative numbers reflecting Jakarta’s larger population. According to guidelines from Cochran⁵⁸, a sample size of around 200 or more is considered sufficient with an acceptable margin of error of 6 to 7%. Moreover, these numbers were higher than in previous studies involving time-use surveys and travel behaviour⁵⁹.

Modelling framework and specifications

Figure 2 illustrates the analytical framework used in this study, which involved three multivariate ordered regression models (MORMs)⁶⁰. The use of a MORM allows for an exploratory analysis of the associations and correlation structures among ordered categorical outcomes in a computationally efficient and interpretable way^{61, 62–63}. This helps in identifying the key relationships, validating dependencies, and refining the model specifications before applying more complex methodologies. In this research, MORMs were used to explore the relationships between activities across locations, timeframes, and purposes (Fig. 2). This modelling framework is well-suited to handling ordinally categorised time-use data while accounting for the correlations among dependent variables, such as different types of activity reported by the same individual. More importantly, a MORM offers a flexible alternative to traditional time-budget models as the fixed-time constraint is relaxed, which is especially relevant in the context of multitasking, where primary and secondary activities may overlap, resulting in reported durations exceeding 24 h. This feature makes a MORM particularly appropriate for modelling time use, where strict additivity assumptions may not hold.

Fig. 2 Modelling framework. [Images not available. See PDF.]

Figure 2 illustrates the analytical framework combining multivariate ordered regression models (MORMs) to analyse the relationships between virtual and physical activities. The model is structured into two components: time allocation and relationships. The framework allows examination of activity patterns across purposes (mandatory, maintenance, leisure), locations (home, work, other), and days (weekdays, Friday, weekends). The figure visualizes the integrated approach used to capture the complex interactions between activities, including multitasking, and highlights how socio-demographic, residential, and personality factors are incorporated into the model.

The dependent variable in these models is the total duration of each activity, for the main/secondary activities and virtual/physical activities. Each MORM is used to analyse the relationships by dividing the dependent variables based on the purpose, location and day. Since ordered data are required for MORMs, the total duration of each type of activity was grouped into several ordered categories. To define the ordinal duration categories used in the MORM analysis, this study referred to the empirical distribution of daily time-use patterns collected from the one-week diary, as shown in the time-use characteristics (see Fig. 3 and Supplementary Tables 1 to 3). Rather than using equal time intervals, this study derived thresholds that reflected meaningful behavioural groupings across the type of activity (mandatory, maintenance, leisure), location, day, and mode of activity (virtual or physical). Each model incorporates socio-demographic and personality characteristics and residential location as explanatory variables to explore the factors that influence individuals’ time allocation. A further explanation of the model specification is given in the next section.

Fig. 3 Respondents’ time-use characteristics based on purpose. [Images not available. See PDF.]

Figure 3 presents time-use patterns based on activity purpose. Most main virtual activities were leisure-related, while the majority of physical activities were mandatory (averaging 78%), such as sleep or work.

Multivariate Ordered Regression Model

A MORM is an extension of the standard ordered regression model (ORM). The motivation for the development of the MORM stemmed from the observation that ordinal responses are often correlated across multiple or repeated measurements⁶⁰. In this study, the hypothesis is that virtual and physical activities are interconnected and associated with each other. To test this hypothesis, three MORMs are developed to estimate (i) purpose, (ii) location, and (iii) day (Fig. 2). There are two components of each MORM in this study, relating to time allocation and relationships. The time allocation component estimates each activity’s duration, similar to a standard ORM, whereas the relationships component covers the correlation between dependent variables. The technical specifications of both components are detailed below.

In a MORM, the observed categories $Y_{\mathit{ij}}$ represent a categorised version of an underlying latent variable Ŷ_ij for an individual i, where i = 1, …, n, and for a dependent variable j. In this study, the dependent variable is the duration of each activity ( $J_i\in \left\{1,\dots ,\mathrm{J}\right\}$ ), which is transformed into a range/ordered category, and can therefore be treated as an ordered response. Although ORMs are designed for ordinal data, the difference between ordinal, continuous and cardinal responses does not affect the modelling process, as demonstrated by Meyer⁶⁴. Singh et al.⁶⁵ also argue that ordered-response systems allow for the use of a general correlation matrix for underlying latent variables, giving a flexible correlation pattern among the observed outcomes. The MORM equation is formulated as follows (Eq. (1)):

1

$${\hat{Y}}_{ij}^{*}={\beta }_{0j}+x_{ij}{\beta }_j+{\in }_{ij}$$ where ${\widehat{Y}}_{ij}^{*}$ is the latent variable corresponding to observed ordinal response $Y_{\mathit{ij}}$ , ${\beta }_{0}j$ is an intercept term, $x_{\mathit{ij}}$ is a p×1 vector of covariates, ${\beta }_j$ is a vector of regression coefficients, and ${ϵ}_{\mathit{ij}}$ is a mean zero error term with distribution function F. In this study, socio-demographic characteristics (i.e. income, gender, age, etc.), residential location, and personality traits are incorporated as the covariates.

The response variables in a MORM follow cumulative link models, assuming that the n individuals are independent. The number of ordered categories, threshold parameters ${\mu }_j$ , and regression coefficients ${\beta }_j$ are allowed to vary across the outcome dimensions j, accounting for potential heterogeneity among response variables. The relationship between the observed variable $Y_{\mathit{ij}}$ (with K categories) and the latent variable ${\hat{Y}}_{ij}^{*}$ is defined as (Eq. (2)):

2

$$P\left(Y_{\mathit{ji}}\le k\right)=F\left({\mathrm{\mu }}_{\mathit{jk}}-x_{\mathit{ij}}{\beta }_j\right)$$ where k (1, …, K) represents an ordered category, ${\mu }_{\mathit{jk}}$ are the threshold parameters for a dependent variable j with constraint ${\mu }_{j0}\le {\mu }_{j1}\le \dots \le {\mu }_{\mathit{jK}}$ , and F is the cumulative distribution function. Binary observations (i.e., two-category ordinal data) are treated as a special case within this framework.

To account for the dependence among different responses, it is assumed that for each individual i, the vector of error terms ( ${ϵ}_{\mathit{ij}}$ ) follows a multivariate normal distribution, consistent with the multivariate ordered probit framework. This structure enables the use of a Gaussian copula, which allows for flexible modelling of the correlations among multiple ordinal responses. The relationships between multiple measurements or outcomes (several dependent variables) can be represented by various correlation structures. These correlation structures are the relationship component that is used in this study to identify the relationships between virtual and physical activities. The approach applied here is similar to that of prior studies that have identified relationships between travel behaviour indicators, such as car ownership⁶² or cycling behaviour⁶³. In this study, an autoregressive correlation (AR) model of order one is employed, which is well-suited to longitudinal data. Given equally spaced time points k and l, the correlation between $Y_{\mathit{jik}}$ and $Y_{\mathit{jil}}$ is ${\rho }_{\mathit{kl}}={\rho }^{\mathit{|k}-l]}$ , where ${\rho }_{\mathit{kl}}$ represents the correlation parameter. This AR structure implies an exponential decay in correlation as the time lag increases. The correlation parameter is modelled as a function of subject-specific covariates $Z_i$ , ensuring that the individual heterogeneity is appropriately captured. The mvord package⁶⁰ in R Studio (where the code is available publicly) is used to estimate the MORM, as it supports the flexible specification of error structures and copula-based multivariate dependencies under the probit assumption. Following the work of Hirk et al.⁶⁰, a hyperbolic tangent transformation is applied to re-parameterise the linear term $Z_i\gamma$ in terms of a correlation parameter, ${\rho }_i=\tanh \left(Z_i\gamma \right)$ , which ensures that the correlation parameter remains within the valid range of (− 1,1).

Results

Respondents’ socio-demographic and personality characteristics

Table 2 summarises the individuals’ socio-demographic characteristics. The respondents in this study were predominantly male, employed and aged 35 or older. They mostly (59%) came from middle-income households (IDR 3 million/USD 189 to IDR 8 million/USD 506). The majority were well-educated, with 53% holding a bachelor’s degree, and 70% lived in households with more than two people. This description indicates a demographic skew in the dataset, with an overrepresentation of individuals aged 20–34 and an underrepresentation of those over 50. This is a common challenge in research in investigating ICT and its impact on travel behaviour, as younger populations are more likely to engage with digital services and app-based platforms^66,67. In this study, eligibility was limited to individuals with experience using smartphones, which may have excluded some older participants. This criterion was necessary in order to meet the study’s objective of examining the impact of ICT usage, but it is worth noting that the findings may not fully capture the experiences of older or less digitally engaged individuals and should be interpreted with this caveat in mind; future data collection that accommodates a broader population is therefore needed.

Table 2. Respondents’ socio-demographic and time use characteristics

Variables	Sample	Populationa	Average Daily Virtual Main Activities	Average Daily Virtual Secondary Activities
			(Minutes)	(Minutes)
Age
< 22 years old	3.0%	10.7%	226.2	265.5
22–30 years old	60.6%	32.9%	193.3	203.5
31–50 years old	31.8%	31.1%	144.0	181.2
>50 years old	4.6%	25.3%	177.3	177.9
Gender
Male	56.2%	50.4%	191.5	190.0
Female	43.8%	49.6%	155.8	202.8
Household monthly incomeǂ
Low-income household (< IDR. 3.000.000)	21.0%	-	160.3	204.6
Middle Income (IDR. 3.000.000 – IDR. 8.000.000)	59.8%	-	169.8	182.7
High Income (> IDR. 8.000.000)	19.2%	-	190.2	204.7
Occupation
Worker	70.1%	-	156.9	190.7
Non-Worker (retired/not work)	11.9%	-	184.1	175.6
Student	18.0%	-	245.5	228.1
Degree of education
Junior/senior high-school	45.1%	-	175.9	198.9
Bachelors’ degree	53.8%	-	162.3	199.8
Master/Doctorals’ degree	1.1%	-	185.9	192.1
Household size
Single	8.1%	-	233.7	134.7
Two members	21.0%	-	133.8	178.4
Three members	31.8%	-	161.0	192.5
Four numbers or more	39.2%	-	198.0	213.5
Residential location
Bandung	28.2%	-	222.6	204.8
Denpasar	22.6%	-	110.1	183.0
Cianjur	18.4%	-	146.3	182.2
Jakarta	30.9%	-	194.1	190.8

^ǂ1 USD = 15,875 IDR as of 8 December 2024; ^abased on aggregated statistics from four cities ^abased on aggregated statistics from four cities^{87, 88, 89–90}.

Table 2 summarizes the age, gender, income, occupation, education, household size, and residential city of the respondents, along with their average daily virtual activity durations (main and secondary). Younger individuals and students engage more in virtual activities, while males and urban dwellers report longer online durations. The sample overrepresents the digitally engaged younger population, common in ICT-related behavioural studies.

Table 3 describes the respondents’ personality traits. In response to the BFI questions in the questionnaire, respondents filled in a self-report five-level Likert scale (1: strongly disagree to 5: strongly agree). The statements in Table 3 reflect the final items that were retained after the factor loadings had been evaluated through confirmatory factor analysis (CFA). Items with loadings of below 0.50 were excluded, to ensure construct validity⁶⁸. The results show that respondents described themselves as sociable (extraversion, 3.9), respectful towards others (agreeableness, 4.06), and persistent (conscientiousness, 4.09). Most lived in areas with good internet access (4.17) and close to activity centres (4.07) such as shopping districts. CFA was used to analyse these personality traits using a predefined framework provided by Soto and John⁵⁷. To assess the suitability of the data for principal component analysis, this study conducted the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy. The KMO values for each personality trait ranged from 0.602 to 0.751 (see Table 3), indicating an acceptable level of sampling adequacy based on Kaiser⁶⁹.

Table 3. Respondents’ personality characteristics

Questions		Mean	Loadingsa
Personality traits. I am a person who…
Extraversion (KMO: 0.745)	tends to be quiet.	2.64	−0.69
	is outgoing or sociable.	3.90	0.79
	is dominant or acts as a leader.	3.30	0.71
	is hoping other to be a leader	3.11	−0.51
	is full of energy.	3.76	0.73
Agreeableness (KMO: 0.644)	is compassionate or has a soft heart.	3.98	0.80
	is respectful or treats others with respect.	4.06	0.84
	assumes the best about people.	3.83	0.73
Conscientiousness (KMO: 0.751)	keeps things neat and tidy.	3.81	−0.71
	tends to be disorganised	2.39	0.60
	is persistent or works until the task is finished.	4.09	−0.65
	has difficulty getting started on tasks.	2.66	0.77
	can be somewhat careless.	3.19	0.61
Negative emotionality (KMO: 0.602)	worries a lot.	3.12	0.60
	tends to feel depressed or blue.	2.41	0.79
	is temperamental or gets emotional easily.	2.33	0.64
Open-Mindedness (KMO: 0.610)	is fascinated by art, music, or literature	3.66	0.51
	is complex or a deep thinker.	3.61	0.80
	is original or comes up with new ideas.	3.51	0.83

^aloadings are calculated based on the CFA of the motivation and personality trait factors. Bartlett scores were generated for each case of each factor. A more positive value of Bartlett score represents a higher representation of the particular personality traits, and vice versa; based on the Bartlett scores for each factor in the CFA.

Table 3 presents the mean scores and factor loadings for selected items from the Big Five Inventory after confirmatory factor analysis (CFA). Traits like Extraversion, Conscientiousness, and Agreeableness showed strong loadings, confirming internal validity. Kaiser-Meyer-Olkin (KMO) values indicate the data were suitable for personality trait extraction and modelling.

Respondents’ Time-use characteristics

The time-use characteristics of the respondents, based on their socio-demographics, are shown to the right of Table 2. Males (202 min) and students (245 min) engaged in main virtual activities more frequently than their counterparts. Respondents from Jakarta (194 min) and Bandung (222 min) reported longer daily durations for primary virtual activities compared to those in Denpasar and Cianjur. Respondents living alone spent the most time on primary virtual activities (233 min), whereas those living in households with more than three people devoted the most time to secondary virtual activities (213 min). Respondents who were highly educated had the highest average participation in main virtual activities (185 min).

The respondents’ time-use characteristics were categorized, as illustrated in Figs. 3 to 5, with a more detailed description provided in the Supplementary Materials (Supplementary Tables 1 to 3). Figure 3 presents time-use patterns based on activity purpose. Most main virtual activities were leisure-related, while most of physical activities were mandatory (averaging 78%), such as sleep or work. Figure 4 displays time-use characteristics based on location. Consistent with prior studies, most respondents reported spending the majority of their time at home (720–1080 min, for more than 68%) and engaged in mandatory activities (720–1080 min, for more than 65%)²⁵. Figure 5 describes time-use characteristics by day of the week. During weekdays, approximately 51–56% of respondents engaged in virtual main activities for 90–360 min. Virtual mandatory tasks were more prominent on weekdays, particularly in Jakarta (16.6% of daily time) and Bandung (18.5%). On weekends, physical leisure activities became more dominant. Overall, due to the advantages of ICT platforms, virtual activities played a greater role in secondary tasks. However, physical activities continued to dominate daily routines, with virtual activities accounting for around 22% of the total duration of primary and secondary activities. Notably, 86–88% of respondents engaged in physical main activities for more than 1080 min.

Fig. 4 Respondents’ time-use characteristics based on location. [Images not available. See PDF.]

Figure 4 displays time-use characteristics based on location. Consistent with prior studies, most respondents reported spending the majority of their time at home (720–1080 min, for more than 68%) and engaged in mandatory activities (720–1080 min, for more than 65%).

Fig. 5 Respondents’ time-use characteristics based on day. [Images not available. See PDF.]

Figure 5 describes time-use characteristics by day of the week. During weekdays, approximately 51–56% of respondents engaged in virtual main activities for 90–360 min Virtual mandatory tasks were more prominent on weekdays, particularly in Jakarta (16.6% of daily time) and Bandung (18.5%). On weekends, physical leisure activities became more dominant.

Estimation of the time allocation component

The estimation results for the time allocation components for the three different models based on purpose, location, and day are presented in Table 4, and the threshold and goodness-of-fit parameters of the models are given in Table 5. In addition to the log-likelihood, AIC, and BIC values, pseudo-R² is included to reflect the explanatory power of each model. The pseudo-R² values are 0.086 for the purpose model, 0.065 for the location model, and 0.058 for the day model. These are modest values, a typical finding for multivariate ordinal models that involve complex behaviours (i.e., daily activity patterns) with extensive explanatory variables. These models are based on categories derived from self-reported seven-day time-use diaries, with 12 dependent variables and 16 explanatory variables in each model, which adds considerable variability. Consequently, it is normal for pseudo-R² values to be lower than for linear regression models with continuous outcomes. Prior studies using behavioural datasets (i.e., time-use and vehicle ownership) have also reported modest pseudo-R² values, typically below 0.1^62,63,70,71.

Table 4. Estimation of the Daily Time Allocation Based on Activity Purposes, Locations and Days

Model 1: Purposes
Variables	Main Activities						Secondary Activities
	VR	VR	VR	PHY	PHY	PHY	VR	VR	VR	PHY	PHY	PHY
	MAN	LEI	MAIN	MAN	LEI	MAIN	MAN	LEI	MAIN	MAN	LEI	MAIN
Socio-demographic
Male (ref.: female)	n.s.	0.19 **	n.s.	n.s.	0.16 *	−0.53 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Age (ref.: > 50 years old)
< 20 years old [D]	0.65 *	0.49 *	n.s.	n.s.	n.s.	n.s.	1.21 **	n.s.	0.79 **	n.s.	n.s.	n.s.
20–34 years old [D]	0.41 *	n.s.	n.s.	n.s.	−0.5 **	n.s.	n.s.	n.s.	n.s.	n.s.	−0.55 **	n.s.
35–49 years old [D]	n.s.	−0.27 *	n.s.	0.44 **	−0.48 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Occupation (ref.: non-workers)
Workers [D]	n.s.	−0.41 **	−0.88 **	1.83 **	−0.42 **	−1.34 **	n.s.	n.s.	−0.76 **	n.s.	0.44 **	n.s.
Students [D]	n.s.	n.s.	−0.55 *	1.32 **	n.s.	−1.52 **	n.s.	0.28 *	−0.52 *	n.s.	0.42 **	n.s.
Household income (ref.: High income)
Low household income [D]	−0.44 **	n.s.	n.s.	n.s.	0.24 **	n.s.	−0.59 **	−0.33 **	n.s.	−0.56 **	−0.25 *	n.s.
Mid household income [D]	n.s.	n.s.	−0.33 *	0.17 *	n.s.	n.s.	−0.53 **	n.s.	n.s.	−0.48 **	0.31 **	n.s.
Household size	−0.07 *	−0.09 **		0.06 *	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Residential location (ref.: Cianjur)
Jakarta [D]	−0.84 **	1.01 **	n.s.	−0.24 *	n.s.	−0.4 **	−0.77 **	−0.46 **	−0.53 **	−0.27 *	1.07 **	n.s.
Bandung [D]	0.31 **	0.62 **	0.83 **	−0.89 **	0.82 **	−1.09 **	−1.67 **	n.s.	−0.32 *	−0.62 **	n.s.	n.s.
Denpasar [D]	n.s.	−0.22 **	n.s.	−0.72 **	1.05 **	n.s.	n.s.	n.s.	n.s.	0.56 **	1.04 **	n.s.
Personality traits
Extraversion	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	−0.08 *	n.s.	n.s.	n.s.	n.s.
Agreeableness	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Conscientiousness	n.s.	−0.13 **	n.s.	0.11 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Negative emotionally	n.s.	0.12 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	−0.13 **	n.s.
Open-Mindedness	n.s.	n.s.	n.s.	n.s.	0.12 **	n.s.	n.s.	0.08 *	n.s.	n.s.	n.s.	n.s.
Model 2: Locations
Variables	Main Activities						Secondary Activities
	VR	VR	VR	PHY	PHY	PHY	VR	VR	VR	PHY	PHY	PHY
	H	W	O	H	W	O	H	W	O	H	W	O
Socio-demographic
n.s.	n.s.	n.s.	n.s.	−0.21 **	−0.2 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Age (ref.: > 50 years old)
< 20 years old [D]	0.54 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	−0.87 **	n.s.
20–34 years old [D]	n.s.	n.s.	n.s.	n.s.	0.35 **	n.s.	n.s.	n.s.	n.s.	−0.44 **	−0.55 **	n.s.
35 - 49 years old [D]	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	−0.29 *	n.s.	n.s.
Occupation (ref.: non-workers)
Workers [D]	−0.47 **	0.97 **	n.s.	−1.49 **	1.77 **	−0.28 **	−0.42 **	1.58 **	n.s.	n.s.	1.42 **	n.s.
Students [D]	n.s.	0.58 **	n.s.	−0.92 **	1.07 **	−0.41 **	n.s.	1.13 **	n.s.	n.s.	1.43 **	n.s.
Household income (ref.: High income)
Low household income [D]	−0.22 *	−0.29 *	0.26 *	n.s.	−0.34 **	0.31 **	−0.42 **	−0.36 **	0.26 *	−0.29 **	−0.55 **	n.s.
Mid household income [D]	−0.24 **	n.s.	0.25 **	n.s.	−0.16 *	0.25 **	−0.21 **	n.s.	0.25 **	0.23 **	n.s.	0.31 **
Household size	−0.09 **	−0.07 *	n.s.	n.s.	0.06 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Residential location (ref.: Cianjur)
Jakarta [D]	0.74 **	0.63 **	−0.75 **	−0.64 **	0.31 **	−0.6 **	−0.42 **	n.s.	−0.75 **	0.79 **	0.58 **	n.s.
Bandung [D]	0.54 **	n.s.	−0.28 **	−0.65 **	n.s.	n.s.	n.s.	−0.4 **	−0.28 **	−0.33 **	−0.32 *	n.s.
Denpasar [D]	−0.23 **	n.s.	n.s.	−0.61 **	0.25 **	n.s.	−0.3 **	n.s.	n.s.	0.61 **	1.01 **	0.53 **
Personality traits
Extraversion	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Agreeableness	n.s.	0.12 **	n.s.	n.s.	0.09 **	n.s.	n.s.	0.09 **	n.s.	n.s.	n.s.	n.s.
Conscientiousness	−0.11 **	n.s.	0.09 *	n.s.	n.s.	n.s.	n.s.	n.s.	0.09 *	n.s.	n.s.	n.s.
Negative emotionally	0.12 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	−0.13 **	n.s.
Open-mindedness	n.s.	−0.12 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Model 3: Days
Variables	Main Activities						Secondary Activities
	VR	VR	VR	PHY	PHY	PHY	VR	VR	VR	PHY	PHY	PHY
	MT	FR	SS	MT	FR	SS	MT	FR	SS	MT	FR	SS
Socio-demographic
Male (ref.: female)	0.18 **	n.s.	0.21 **	−0.34 **	n.s.	-0.32 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Age (ref.: > 50 years old)
< 20 years old [D]	0.68 **	n.s.	n.s.	−0.75 *	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	−0.71 **	n.s.
20–34 years old [D]	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	−0.35 **	−0.45 **	−0.56 **
35–49 years old [D]	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	-0.33 **
Occupation (ref.: non-workers)
Workers [D]	−0.41 **	−0.38 **	n.s.	0.78 **	0.4 *	n.s.	n.s.	n.s.	n.s.	0.4 **	0.41 **	0.37 **
Students [D]	n.s.	n.s.	0.34 **	n.s.	n.s.	n.s.	n.s.	0.28 *	n.s.	0.42 **	0.55 **	0.37 **
Household income (ref.: High income)
Low household income [D]	−0.21 *	−0.35 **	n.s.	0.44 **	0.48 **	n.s.	−0.5 **	−0.37 **	−0.28 **	−0.34 **	−0.27 **	−0.34 **
Mid household income [D]	−0.21 **	−0.18 **	−0.15 *	0.37 **	0.34 **	n.s.	−0.24 **	−0.23 **	n.s.	0.17 *	0.19 **	0.25 **
Household size	−0.1 **	−0.09 **	−0.08 **	0.11 **	0.1 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Residential location (ref.: Cianjur)
Jakarta [D]	0.69 **	0.61 **	0.71 **	n.s.	n.s.	n.s.	−0.57 **	−0.31 **	−0.54 **	0.79 **	0.82 **	0.75 **
Bandung [D]	0.52 **	0.54 **	0.63 **	−0.61 **	−0.73 **	−1.04 **	−0.32 **	n.s.	n.s.	n.s.	−0.41 **	n.s.
Denpasar [D]	n.s.	n.s.	−0.27 **	n.s.	n.s.	n.s.	−0.24 **	n.s.	−0.21 *	n.s.	0.86 **	0.88 **
Personality traits
Extraversion	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Agreeableness	n.s.	−0.07 *	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	0.07 *	n.s.	n.s.	n.s.
Conscientiousness	−0.14 **	−0.13 **	−0.11 **	0.16 **	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Negative emotionally	0.13 **	0.13 **	0.1 **	−0.18 **	n.s.	−0.16 *	n.s.	n.s.	n.s.	n.s.	−0.08 **	n.s.
Open-mindedness	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	0.07 *	n.s.	n.s.	n.s.	n.s.	n.s.

VR Virtual activities, PHY Physical activities, MAN Mandatory activities, LEI Leisure activities, MAIN Maintenance activities, H Home, W Work/school, O Other location, MT Monday to Thursday, FR Friday, SS Saturday to Sunday, *Significant at 10%; ** Significant at 5%, n.s. not significant

Table 4 reports result from three multivariate ordered regression models (MORMs) estimating time allocation based on purpose, location, and day, across virtual/physical and main/secondary activities. It shows how socio-demographics and personality traits influence the amount of time spent on each activity type. Significant patterns include increased virtual engagement among younger, wealthier individuals and greater physical engagement among those in larger households or less urban areas.

Table 5. Thresholds and Goodness-of-fit Parameters

Model 1: Purposes
Variables	Main Activities						Secondary Activities
	VR	VR	VR	VR	VR	VR	VR	VR	VR	VR	VR	VR
	MAN	LEI	MAN	LEI	MAN	LEI	MAN	LEI	MAN	LEI	MAN	LEI
Thresholds:
µ1	0.6 **	−1.29 **	1.02 **	0.45 *	−3.41 **	−3.12 **	1.58 **	−3.05 **	0.58 *	n.s.	0.43 *	1.78 **
µ2	0.92 **	−1.01 **	2.35 **	2.77 **	−2.76 **	−2.49 **	1.9 **	−1.96 **	1.94 **	0.89 **	1.09 **	2.61 **
µ3	1.28 **	−0.54 **	2.89 **		−1.89 **	−1.81 **	2.21 **	−1.01 **	2.8 **	1.88 **	1.94 **	3.34 **
µ4	1.87 **	n.s.			−0.48 **	−1.02 **	3.05 **	n.s.			3.01 **
µ5	3.16 **	1.78 **			1.18 **	n.s.		1.21 **			3.98 **
µ6		3.48 **			3.06 **			2.83 **
Log-likelihood; BIC; AIC; Pseudo-r2	−128195.08; 263961.46; 283208.55; 0.086
Model 2: Locations
Variables	Main Activities						Secondary Activities
	VR	VR	VR	VR	VR	VR	VR	VR	VR	VR	VR	VR
	H	W	H	W	H	W	H	W	H	W	H	W
Thresholds:
µ1	−1.5 **	1.47 **	0.66 **	−5.11 **	1.15 **	−1.11 **	−2.74 **	1.08 **	0.66 **	n.s.	1.66 **	1.48 **
µ2	−1.13 **	2.44 **	1.29 **	−4.53 **	1.29 **	−0.38 *	−1.85 **	1.92 **	1.29 **	0.55 **	2.71 **	2.03 **
µ3	−0.69 **	3.25 **	1.88 **	−4.5 **	1.4 **	n.s.	−0.97 **	2.62 **	1.88 **	1.55 **	3.47 **	2.46 **
µ4	n.s.	3.58 **	2.37 **	−2.99 **	1.58 **	0.9 **	n.s.	3.23 **	2.37 **	2.61 **	3.96 **	2.99 **
µ5	1.22 **		3.42 **	−0.63 **	2.61 **	1.28 **	1.13 **	4.24 **	3.42 **
µ6	3.06 **					2.14 **
µ7						2.49 **
Log-likelihood; BIC; AIC; Pseudo-r2	−163526.25; 334848.17; 354665.65; 0.065
Model 3: Days
Variables	Main Activities						Secondary Activities
	VR	VR	VR	VR	VR	VR	VR	VR	VR	VR	VR	VR
	MT	FR	MT	FR	MT	FR	MT	FR	MT	FR	MT	FR
Thresholds:
µ1	−1.3 **	−1.17 **	−0.97 **	−2.95 **	−2.68 **	−4.22 **	−3.43 **	−2.39 **	−2.72 **	n.s.	n.s.	n.s.
µ2	−1.07 **	−1.02 **	−0.74 **	−0.79 **	−1.08 **	−1.89 **	−2.38 **	−1.58 **	−1.72 **	0.72 **	0.71 **	0.42 *
µ3	−0.64 **	-0.68 **	−0.37 *				−1.38 **	−0.8 **	−0.86 **	1.41 **	1.44 **	1.04 **
µ4	n.s.	0.06	0.41 *				n.s.	n.s.	n.s.	2.51 **	2.2 **	2.07 **
µ5	1.23 **	0.97 **	1.48 **				0.77 **	1.08 **	1.1 **	3.47 **	3.13 **	2.82 **
µ6	3.15 **	2.39 **	3.49 **				2.65 **	2.41 **	2.64 **			3.53 **
Log-likelihood; BIC; AIC; Pseudo-r2	−176800.42; 361286.12; 380822.97; 0.058

VR Virtual activities, PHY Physical activities, MAN Mandatory activities, LEI Leisure activities, MAIN Maintenance activities, H Home, W Work/school, O Other location, MT Monday to Thursday, FR Friday, SS Saturday to Sunday; *Significant at 10%; ** Significant at 5%; n.s. not significant.

This table provides the threshold values, log-likelihood, AIC/BIC, and pseudo-R² for each MORM. The models achieved modest fit (pseudo-R² < 0.1), consistent with complex behavioural data. Thresholds help classify ordinal activity durations, while model diagnostics confirm the appropriateness of using MORM for time-use patterns.

The first part of Table 4 describes the estimation for the time allocation model based on purpose. It is found that younger people are likely to allocate more time to virtual activities for various purposes (mandatory, maintenance, and leisure). Wealthy people also tend to allocate more time to mandatory virtual activities as both main and secondary activities, indicating their greater work-related multitasking behaviour. Students and workers are more likely to allocate more time to physical mandatory activities, and likewise, respondents who live with more people tend to allocate less time to virtual activities and are more likely to allocate time to the main physical mandatory activities.

In regard to residential locations, people who live in larger cities (Jakarta and Bandung) are more likely to allocate more time to virtual activities, especially mandatory and leisure activities, while those who live in a bigger city (Jakarta) tend to allocate more time to secondary virtual activities, especially for mandatory purposes. In terms of personality, people whose emotions fluctuate are more likely to allocate time to virtual leisure activities. Likewise, people who are more creative and open to new things also tend to allocate more time to leisure activities, including physical engagement, and allocate marginally more virtual leisure activities as secondary activities.

The estimation for the time allocation model based on location is described in the second part of Table 4. Female respondents tend to allocate more time to activities in their homes, and especially physical activities. Wealthy people are more likely to allocate more time to virtual activities, as both main and secondary activities, in homes or workplaces. Younger people (less than 20 years old) tend to allocate more time to main virtual activities at home, while students at university or people of early working age (20–34) tend to allocate more time to workplaces or study locations for their main physical activities. Workers are more likely to allocate more time to their workplaces for both virtual and physical activities, as well as main and secondary activities. This study also found that non-workers tend to allocate more time to primary virtual activities at home. Respondents who live with more individuals tend to allocate less time to virtual activities at home and the workplace, but are more likely to allocate more time to physical activities in their workplaces. In regard to residential location, people who live in larger cities (Jakarta and Bandung) tend to allocate more virtual activity to their homes and less to other locations. Individuals who live in Cianjur are likely to perform more physical activities in their homes than people in other cities. One interesting finding is that people in Cianjur also multitask more in their homes. People with kinder personalities tend to allocate more time to virtual and physical activities in their workplaces.

The last part of Table 4 describes the time allocation component based on days. Male respondents are likely to allocate more time to virtual main activities both on weekdays and weekends, while female respondents tend to allocate more time to physical activities throughout the entire week. Wealthy people are more likely to allocate more time to virtual activities during weekdays, but less time to physical activities. They are also more likely to multitask during weekends and weekdays. Younger people (less than 20 years old) tend to allocate more time to main virtual activities but less to physical activities on weekdays, whereas workers tend to allocate more time to physical activities on weekdays. An unexpected finding is that users in Denpasar tend to allocate less time to virtual activities on weekends than people in other cities, whereas individuals living in larger cities (Jakarta and Bandung) tend to allocate time to main virtual activities on both weekdays and weekends. People who are more disorganised are likely to allocate more time to main physical activities on earlier weekdays (Monday to Thursday) and less to main virtual activities on both weekdays and weekends.

Estimation of the relationships component

The relationships component of the models is shown in Table 6. In this table, the relationship is identified using a correlation structure derived from the model’s error structure. Given this correlation, these should not be interpreted as causal effects, but rather as indications of association between patterns of engagement in virtual and physical activities. Table 6 shows the three relationship components as outlined in Fig. 2. The first section of Table 6 summarises the relationships between activities based on their purpose. The model reveals that mandatory physical activities have negative correlations with various virtual activities, indicating an inverse relationship that may reflect substitution patterns. Similarly, physical leisure activities show negative correlations with virtual leisure activities, although the substitution effect is less pronounced than for the mandatory or maintenance physical activities. A higher likelihood of engaging in secondary virtual activities is observed when individuals are involved in main virtual mandatory activities, but no significant inverse relationship (trade-off) between virtual activities based on purpose was found.

Table 6. Estimation of the Relationships Between Activities Based on Activity Purposes, Locations and Days

Error Structure	Main Activities						Secondary Activities
	VR	VR	VR	PHY	PHY	PHY	VR	VR	VR	PHY	PHY	PHY
Purposes	MAN	LEI	MAIN	MAN	LEI	MAIN	MAN	LEI	MAIN	MAN	LEI	MAIN
Main-Virtual-Mandatory	1.00
Main-Virtual-Leisure	0.28	1.00
Main-Virtual-Maintenance	0.37	0.20	1.00
Main-Physical-Mandatory	−0.54	−0.43	−0.40	1.00
Main-Physical-Leisure	0.01	−0.21	0.10	−0.42	1.00
Main-Physical-Maintenance	−0.20	−0.35	−0.34	−0.15	−0.15	1.00
Sec-Virtual-Mandatory	0.18	−0.09	−0.07	0.05	0.01	0.00	1.00
Sec-Virtual-Leisure	0.16	0.01	−0.11	−0.06	0.00	0.12	0.01	1.00
Sec-Virtual-Maintenance	0.23	0.06	0.12	−0.27	0.11	0.03	0.23	0.19	1.00
Sec-Physical-Mandatory	−0.13	0.10	−0.14	−0.08	0.02	0.05	0.14	0.35	0.06	1.00
Sec-Physical-Leisure	−0.06	−0.10	−0.17	0.14	−0.08	0.14	0.12	0.23	−0.11	0.30	1.00
Sec-Physical-Maintenance	0.06	−0.13	0.00	−0.02	−0.03	0.24	0.02	0.23	0.14	0.35	0.32	1.00
Locations	H	W	O	H	W	O	H	W	O	H	W	O
Main-Virtual-Home	1.00
Main-Virtual-Work/School	−0.04	1.00
Main-Virtual-Other	−0.15	−0.29	1.00
Main-Physical-Home	−0.23	−0.13	−0.39	1.00
Main-Physical-Work/School	−0.31	0.36	−0.44	-0.25	1.00
Main-Physical-Other	−0.20	−0.41	0.83	−0.36	−0.34	1.00
Sec-Virtual-Home	0.22	−0.23	0.17	0.18	−0.27	−0.08	1.00
Sec-Virtual-Work/School	−0.29	0.48	−0.21	−0.23	0.73	−0.32	−0.02	1.00
Sec-Virtual-Other	−0.15	−0.29	1.00	−0.39	−0.44	0.83	0.16	−0.21	1.00
Sec-Physical-Home	0.02	−0.02	0.17	0.00	−0.20	0.05	0.32	0.02	0.17	1.00
Sec-Physical-Work/School	−0.16	0.40	−0.16	−0.18	0.53	−0.23	0.05	0.51	−0.16	0.32	1.00
Sec-Physical-Other	−0.05	−0.20	0.65	−0.41	−0.38	0.69	−0.01	−0.27	0.66	0.40	0.01	1.00
Days	MT	FR	SS	MT	FR	SS	MT	FR	SS	MT	FR	SS
Main-Virtual-Mon-Thurs	1.00
Main-Virtual-Friday	0.82	1.00
Main-Virtual-Sat-Sun	0.74	0.67	1.00
Main-Physical-Mon-Thurs	−0.98	−0.84	−0.67	1.00
Main-Physical-Friday	−0.82	−0.99	−0.61	0.86	1.00
Main-Physical-Sat-Sun	−0.67	−0.62	−0.98	0.61	0.57	1.00
Sec-Virtual-Mon-Thurs	0.05	0.02	−0.05	−0.12	−0.12	−0.10	1.00
Sec-Virtual-Friday	−0.01	−0.02	−0.09	−0.05	−0.07	−0.06	0.79	1.00
Sec-Virtual-Sat-Sun	0.01	−0.01	0.01	−0.04	−0.04	−0.16	0.77	0.68	1.00
Sec-Physical-Mon-Thurs	−0.01	−0.02	−0.04	0.10	0.05	0.00	0.26	0.24	0.29	1.00
Sec-Physical-Friday	−0.02	−0.02	−0.03	0.14	0.06	−0.02	0.22	0.21	0.28	0.82	1.00
Sec-Physical-Sat-Sun	−0.03	−0.05	−0.04	0.12	0.09	0.01	0.26	0.27	0.26	0.82	0.77	1.00

Table 6 presents correlation matrices capturing the associations between virtual and physical activities across purposes, locations, and days. It reveals significant substitution effects (e.g., virtual work displacing physical tasks) and complementarity (e.g., multitasking virtual leisure alongside physical activities). Correlations vary by day and location, with weekday virtual activity negatively associated with physical activity, especially for mandatory tasks.

VR Virtual activities, PHY Physical activities, MAN Mandatory activities, LEI Leisure activities, MAIN Maintenance activities, H Home, W Work/school, O Other location, MT Monday to Thursday, FR Friday, SS Saturday to Sunday, Bold value means significant at 5%; black colour means positive correlation; red colour means negative correlation.

The second section of Table 6 shows relationships based on activity location. Physical activities conducted at home exhibit higher correlations with virtual and physical activities in other locations, suggesting greater adaptability. Main and secondary virtual activities conducted in the workplace complement the main physical activities in the same location. However, an inverse relationship (negative correlation) is observed between virtual activities conducted in the workplace and those at home or in other locations, potentially reflecting the flexibility introduced by remote working options. In addition, higher levels of secondary virtual activities were noted in the workplace.

The relationships between activities based on days are also reported in Table 6. Longer durations of main or secondary virtual or physical activities on a given day are positively correlated with longer durations of similar activities on other days. Higher engagement in virtual activities during weekdays is associated with lower levels of physical activity. However, the respondent to the study exhibits higher levels of physical activity at weekends compared to weekdays. The result also notes that the negative correlation between virtual and physical activities is weaker during weekends.

Discussion

Many studies have investigated the ways in which individuals adjust their activities with the help of ICTs in order to engage in tasks virtually. This study sought to illuminate how people allocate their time on a daily basis, and to identify relationships based on activity types (primary and secondary) and modes of engagement (virtual and physical) across various locations, days, and purposes. By analysing longitudinal data from a one-week time-use diary completed by smartphone users in four cities in Indonesia, and three MORMs, this study revealed significant insights regarding the relationships between virtual and physical activities.

The time allocation analysis found that younger and wealthier individuals tend to allocate more time to virtual activities across different locations, especially during weekdays. This finding aligns with prior research indicating that the adoption of ICT activities is higher among wealthier and younger individuals due to enhanced accessibility to advanced ICT, distinct job characteristics, and the complexities of participation in daily activities. However, wealthy individuals tend to spend less time on virtual activities on weekends. In addition, male respondents exhibited higher engagement in daily virtual activities both on weekdays and at weekends, which is a reasonable finding in the Indonesian context, where male employees dominate the workforce. Their utilisation of various apps may be driven by the diverse and complex obligations arising in their workplaces⁷². Moreover, people who live with multiple family members tend to allocate more time to physical activities, especially mandatory ones. This finding may be related to the more complex obligations and relationships between household members⁷³, where a greater number of intra-household activities might require more physical activities. With their varying characteristics and complex obligations across members^74,75, households continue to play a significant role in shaping activity schedules in the digital era. It is interesting to note that while workers devote more time to physical mandatory activities, they also optimise their schedules by engaging in virtual secondary activities, particularly for maintenance tasks such as shopping. This behaviour is likely to reflect their response to diverse work demands, as they multitask to address needs in different areas of their lives. Personality traits are also found to shape time allocation. People whose emotions fluctuate or who are more creative tend to allocate more time to leisure activities, but with physical engagement and multitasking. Individuals who are kinder and more sociable tend to allocate more time to various virtual and physical activities, especially in their workplaces. These findings suggest that people with different personalities allocate time differently due to their psychological constructs.

Variations in time allocations were also found based on residential location. People living in Denpasar, a city famous for tourists and traditional culture, tend to engage in more physical activities on weekends, while residents of Jakarta and Bandung, which are more metropolitan, tend to allocate more time to virtual activities, especially for leisure purposes. Residents of Jakarta and Bandung also allocate more time to their workplaces. These distinctions could be attributed to the variations in socioeconomic characteristics and cultural settings, diversity of activities, and flexible time–space priming in these urban areas. Jakarta is characterised by a more intensely digitalised working culture, while heavy congestion leads to more time for commuting, leaving less time for performing daily activities^25,55. In contrast, Denpasar offers a more vibrant cultural environment, where people frequently engage in religious activities together outside their homes, often gathering at village landmarks known as banjar⁷⁶. There are over 750 banjar in Denpasar serving as hubs for socio-cultural activities, and the local government strategically supports and enhances their role⁷⁷. This effort extends beyond preserving cultural heritage, with the aim of leveraging these spaces for economic growth and tourism development^78,79. This communal practice may contribute to increased physical activity, underlining the unique impact of cultural settings on travel behaviour and activity participation.

Our analysis reveals negative correlations between physical and virtual activities across various types of activity, including mandatory, maintenance, and leisure, indicating patterns consistent with substitution. This relationship between virtual and physical activities is complex, and may be influenced by factors such as congestion and air pollution in Indonesian cities⁸⁰. However, physical leisure activities have a positive association with their virtual alternative, due to their unique nature. Previous studies have highlighted the limited adaptability of leisure activities to virtual platforms, as many require physical interaction⁸¹. Main virtual activities are positively associated with main physical activities and secondary virtual activities. This finding supports the role of multitasking in maximising physical activities, as noted in previous studies¹⁷. Notably, secondary virtual activities are positively correlated with work-related virtual tasks, particularly those related to maintenance, consistent with findings suggesting that workers often multitask to address various needs and to use their time more efficiently.

The findings also reveal an association between activity locations. Workplace tasks have a negative correlation with other locations, such home or in other locations, indicating a degree of flexibility. This reflects the growing prevalence of teleworking and likely to be driven by advancements in internet quality and the proliferation of teleworking spaces in urban areas following COVID-19. During the survey period, hybrid working options were offered by the Indonesian government and businesses, which further enhanced the attractiveness of teleworking^82,83. Virtual activity patterns also differ by day. Additionally, virtual activity during the week negatively associated with physical activity, although this inverse relationship is less evident at weekends, suggesting a more relaxed correlation between virtual and physical activities during this period. Differences in time allocation and structuring on weekdays versus weekends, along with unobserved factors such as fatigue from intensive engagement in virtual activities, may influence these patterns. This trend aligns with the possibility of a “virtual activity budget,” akin to travel and out-of-home activity time budgets⁸⁴. For example, individuals who work and commute extensively during the week may prioritise rest or physical activities on weekends. Nevertheless, further investigation into the causal relationships between virtual and physical activities is necessary, representing a valuable opportunity for future research.

The high level of smartphone use in Indonesia is expected to increase due to advancements in the ICT infrastructure and the digital economy, leading to more virtual activities. The government also plans to intensify digitalisation, with the aim of maximising the efficiencies arising from ICT^85,86. An understanding of the relationships between virtual and physical activities is therefore essential in order to foster the transition towards more sustainable cities. Based on the findings of this study, several policy recommendations can be proposed.

Firstly, a negative correlation was observed between activities in workplaces and homes or other locations, while a stronger positive correlation was identified between activities in workplaces and venues such as cafés and restaurants. This suggests that while teleworking may reduce reliance on traditional workplaces, it often involves travel rather than remaining at home. Consequently, the potential reduction in travel distances due to teleworking may be less significant than is often assumed. These findings highlight the need for urban planning and digitalisation to evolve together, and hence, a need to revisit our transport model. Urban and transport planners should integrate these insights into digital behaviour and recognise that virtual activity patterns (e.g., remote work, e-shopping) influence the spatial and temporal demand for physical infrastructure. In the short term, governments, businesses, and educational institutions implementing work-from-home policies should consider designing teleworking spaces near residential areas, to minimise travel distances. This can be done by ensuring that distributed teleworking hubs or co-working spaces are located in residential neighbourhoods. Short-term strategies and long-term planning for ICT use and urban planning, particularly in terms of facilitating flexible work and e-commerce, may reduce unnecessary long-distance trips and lower emissions in order to reach environmental goals.

Secondly, despite the findings on the relationship between virtual and physical activities, the results also show that engagement in virtual activities varies across income, household characteristics, and city size. This indicates that the benefits of digitalisation may not be evenly distributed, which may potentially exacerbate social and spatial inequalities. Urban planners and policymakers should recognise this disparity and prioritise the development of digital equity policies, which can help to ensure that marginalised and underserved populations and regions are not excluded from the time- and resource-saving opportunities offered by digitalisation. This can be done by improvements to both infrastructure and personal capacity, which will involve educational and infrastructure-related institutions. Equitable access to ICT infrastructure, digital literacy initiatives, and inclusive service design are essential to ensure that digitalisation contributes to social inclusion and balanced urban development.

Thirdly, the analysis revealed a negative correlation between virtual and physical activities throughout the week, which becomes weaker when weekdays are compared to weekends. These findings suggest variations in time allocation patterns, with individuals engaging more in physical activities at weekends. Broader digitalisation to support virtual activities on weekdays may be associated with increased physical activity and travel on weekends, which policymakers and transport planners should take into account. This is particularly relevant for larger cities such as Jakarta and Bandung, which have denser facilities and more intense economic activity compared to smaller cities such as Denpasar, where cultural practices and infrastructure already encourage sustainable out-of-home activities.

Finally, it is essential to account for the growing trend of virtual maintenance activities such as e-shopping, which are often performed as both primary and secondary tasks, particularly during work hours. Synergy between transport and urban planning can be exploited to facilitate these trends by encouraging sustainable transport options for these services, or by situating shopping facilities closer to activity centres or residential areas. However, further analysis is necessary to break down activity data into specific categories, such as the type of maintenance activity (e.g., food and beverage, groceries, or clothing) and the type of delivery service (e.g., same-day or next-day delivery), to better identify optimal locations for improving e-shopping services.

Despite these findings, there are several limitations to this research that can offer avenues for future work. The data collected for this study focused on primary and secondary activities, and tertiary activities and identification of activity fragmentation were neglected. In addition, the dataset may have suffered from sample bias toward younger, more digitally literate individuals, with the exclusion of older or less tech-savvy populations. A more comprehensive dataset that captures tertiary activities and activity fragmentation and which represents broader populations proportionally could enable new, detailed research into how people arrange their activities in this digital age. Moreover, the measurement of spatial attributes in this study is represented only by an aggregated city-level variable, and the details of spatial accessibility and the built environment are not considered. Extending the modelling framework to account for these variables could enhance the current understanding of how these factors influence interactions between activities. Gaining a deeper understanding of how these specific ICT services and their detailed spatial characteristics impact travel demand and activity participation could offer valuable insights for the development of policies to address their societal and environmental implications. Moreover, although this study emphasised the role of virtual activities in daily life, the rebound effects of particular ICT services, such as e-shopping, ride-hailing, and telecommuting, on activity participation and travel behaviour require further investigation.

Acknowledgements

The authors thank all the participants and interviewers of the survey. This study is funded by the Research Grant from ITENAS Bandung (Hibah Penelitian Utama), the WCR Grant from The Ministry of Research, Technology, and Higher Education, the Republic of Indonesia, and the DAVeMoS BMK Endowed Professorship in Digitalisation and Automation in Transport Systems (FFG project number: 862678). The first author also acknowledged the Ernst Mach Grant ASEA-UNINET for the doctoral scholarship.

Author contributions

The authors confirm contributions to the paper as follows: study conception and design: M.R.; Y.O.S.; data collection: M.R.; analysis: M.R.; interpretation of results: M.R.; Y.O.S.; T.B.J.; S.S.; supervision: Y.O.S.; T.B.J.; S.S.; funding acquisition: M.R.; Y.O.S.; T.B.J.; first draft manuscript: M.R.; manuscript editing: Y.O.S.; T.B.J.; S.S. All authors reviewed the results and approved the final version of the manuscript.

Data availability

Data collected for this research cannot be shared publicly due to an agreement with the funder.

Code availability

This study uses mvord package from R Studio.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s44333-025-00065-1.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Dharmowijoyo, DBE; Susilo, YO; Karlström, A. On complexity and variability of individuals’ discretionary activities. Transportation; 2018; 45, pp. 177-204. [DOI: https://dx.doi.org/10.1007/s11116-016-9731-5]

2. Susilo, YO; Liu, C. Examining the relationships between individual’s time use and activity participations with their health indicators. Eur. Transp. Res. Rev.; 2017; 9, [DOI: https://dx.doi.org/10.1007/s12544-017-0243-y] 26.

3. Srinivasan, S; Bhat, CR. An exploratory analysis of joint-activity participation characteristics using the American time use survey. Transportation; 2008; 35, pp. 301-327. [DOI: https://dx.doi.org/10.1007/s11116-007-9155-3]

4. Hägerstrand, T. What about people in Regional Science?. Pap. Reg. Sci. Assoc.; 1970; 24, pp. 6-21. [DOI: https://dx.doi.org/10.1007/BF01936872]

5. Schwanen, T; Kwan, M-P. The Internet, mobile phone and space-time constraints. Geoforum; 2008; 39, pp. 1362-1377. [DOI: https://dx.doi.org/10.1016/j.geoforum.2007.11.005]

6. Mokhtarian, P. L. & Tal, G. Impacts of Ict on Travel Behavior: A Tapestry of Relationships. SAGE Handb. Transp. Stud. 241–260 (2013).

7. Shi, K; De Vos, J; Yang, Y; Witlox, F. Does e-shopping replace shopping trips? Empirical evidence from Chengdu, China. Transp. Res. Part Policy Pract.; 2019; 122, pp. 21-33. [DOI: https://dx.doi.org/10.1016/j.tra.2019.01.027]

8. Joewono, TB; Rizki, M; Belgiawan, PF; Irawan, MZ. Why do shoppers keep making online shopping trips? Learning from evidence in Bandung, Indonesia. Asian Transp. Stud.; 2020; 6, 100016. [DOI: https://dx.doi.org/10.1016/j.eastsj.2020.100016]

9. Cao, X. The relationships between e-shopping and store shopping in the shopping process of search goods. Transp. Res. Part Policy Pract.; 2012; 46, pp. 993-1002. [DOI: https://dx.doi.org/10.1016/j.tra.2012.04.007]

10. Melo, PC; de Abreu e Silva, J. Home telework and household commuting patterns in Great Britain. Transp. Res. Part Policy Pract.; 2017; 103, pp. 1-24. [DOI: https://dx.doi.org/10.1016/j.tra.2017.05.011]

11. Asgari, H; Jin, X. Impacts of Telecommuting on Nonmandatory Activity Participation: Role of Endogeneity. Transp. Res. Rec.; 2017; 2666, pp. 47-57. [DOI: https://dx.doi.org/10.3141/2666-06]

12. Farag, S; Dijst, M; Lanzendorf, M. Exploring the Use of E-Shopping and Its Impact on Personal Travel Behavior in the Netherlands. Transp. Res. Rec.; 2003; 1858, pp. 47-54. [DOI: https://dx.doi.org/10.3141/1858-07]

13. Pabilonia, S. W. & Vernon, V. Telework and Time Use. in Handbook of Labor, Human Resources and Population Economics (ed. Zimmermann, K. F.) 1–26 (Springer International Publishing, Cham, 2020). https://doi.org/10.1007/978-3-319-57365-6_274-1.

14. Xu, L; Saphores, J-D. Does e-shopping impact household travel? Evidence from the 2017 U.S. NHTS. J. Transp. Geogr.; 2024; 115, 103827. [DOI: https://dx.doi.org/10.1016/j.jtrangeo.2024.103827]

15. Hostettler Macias, L; Ravalet, E; Rérat, P. Potential rebound effects of teleworking on residential and daily mobility. Geogr. Compass; 2022; 16, [DOI: https://dx.doi.org/10.1111/gec3.12657] e12657.

16. Pelchen, L. Internet Usage Statistics In 2024. Forbes Homehttps://www.forbes.com/home-improvement/internet/internet-statistics/ (2024).

17. Kenyon, S. Internet Use and Time Use: The importance of multitasking. Time Soc; 2008; 17, pp. 283-318. [DOI: https://dx.doi.org/10.1177/0961463X08093426]

18. Liu, C; Susilo, YO; Karlström, A. Jointly modelling individual’s daily activity-travel time use and mode share by a nested multivariate Tobit model system. Transp. Transp. Sci.; 2017; 13, pp. 491-518.

19. Bin, E; Andruetto, C; Susilo, Y; Pernestål, A. The trade-off behaviours between virtual and physical activities during the first wave of the COVID-19 pandemic period. Eur. Transp. Res. Rev.; 2021; 13, [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38624632][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7872826][DOI: https://dx.doi.org/10.1186/s12544-021-00473-7] 14.

20. Susilo, Y. O. & Liu, C. Exploring patterns of time-use allocation and immobility behaviours in the Bandung Metropolitan Area, Indonesia. in Urban Mobilities in the Global South (Routledge, 2017).

21. Kitamura, R; Fujii, S; Pas, EI. Time-use data, analysis and modeling: toward the next generation of transportation planning methodologies. Transp. Policy; 1997; 4, pp. 225-235. [DOI: https://dx.doi.org/10.1016/S0967-070X(97)00018-8]

22. Krueger, AB et al. Time Use and Subjective Well-Being in France and the U.S. Soc. Indic. Res.; 2009; 93, pp. 7-18. [DOI: https://dx.doi.org/10.1007/s11205-008-9415-4]

23. Hilgert, T; von Behren, S; Eisenmann, C; Vortisch, P. Are Activity Patterns Stable or Variable? Analysis of Three-Year Panel Data. Transp. Res. Rec.; 2018; 2672, pp. 46-56. [DOI: https://dx.doi.org/10.1177/0361198118773557]

24. Susilo, Y. O. The Short-term Variability and the Long-term Changes of Individual Spatial Behavior in Urban Areas. (京都大学 (Kyoto University), 2005).

25. Dharmowijoyo, DBE; Susilo, YO; Karlström, A; Adiredja, LS. Collecting a multi-dimensional three-weeks household time-use and activity diary in the Bandung Metropolitan Area, Indonesia. Transp. Res. Part Policy Pract.; 2015; 80, pp. 231-246. [DOI: https://dx.doi.org/10.1016/j.tra.2015.08.001]

26. Hafezi, MH; Liu, L; Millward, H. A time-use activity-pattern recognition model for activity-based travel demand modeling. Transportation; 2019; 46, pp. 1369-1394. [DOI: https://dx.doi.org/10.1007/s11116-017-9840-9]

27. Susilo, YO; Kitamura, R. Analysis of Day-to-Day Variability in an Individual’s Action Space: Exploration of 6-Week Mobidrive Travel Diary Data. Transp. Res. Rec.; 2005; 1902, pp. 124-133. [DOI: https://dx.doi.org/10.1177/0361198105190200115]

28. Sugie, Y; Zhang, J; Fujiwara, A. A weekend shopping activity participation model dependent on weekday shopping behavior. J. Retail. Consum. Serv.; 2003; 10, pp. 335-343. [DOI: https://dx.doi.org/10.1016/S0969-6989(02)00053-X]

29. Yamamoto, T; Kitamura, R. An analysis of time allocation to in-home and out-of-home discretionary activities across working days and non- working days. Transportation; 1999; 26, pp. 231-250. [DOI: https://dx.doi.org/10.1023/A:1005167311075]

30. Lu, X; Pas, EI. Socio-demographics, activity participation and travel behavior. Transp. Res. Part Policy Pract.; 1999; 33, pp. 1-18. [DOI: https://dx.doi.org/10.1016/S0965-8564(98)00020-2]

31. Chen, X; Kemperman, A; Timmermans, H. Socio-demographics, neighborhood characteristics, time use, and leisure-time physical activity engagement patterns over the life course. SSM - Popul. Health; 2022; 19, 101244. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36203469][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9529596][DOI: https://dx.doi.org/10.1016/j.ssmph.2022.101244]

32. Raux, C et al. Travel and activity time allocation: An empirical comparison between eight cities in Europe. Transp. Policy; 2011; 18, pp. 401-412. [DOI: https://dx.doi.org/10.1016/j.tranpol.2010.11.004]

33. Lee, Y; Hickman, M; Washington, S. Household type and structure, time-use pattern, and trip-chaining behavior. Transp. Res. Part Policy Pract.; 2007; 41, pp. 1004-1020. [DOI: https://dx.doi.org/10.1016/j.tra.2007.06.007]

34. Zhang, J; Timmermans, HJP; Borgers, A. A model of household task allocation and time use. Transp. Res. Part B Methodol.; 2005; 39, pp. 81-95. [DOI: https://dx.doi.org/10.1016/j.trb.2004.03.001]

35. Kang, H; Scott, DM. Exploring day-to-day variability in time use for household members. Transp. Res. Part Policy Pract.; 2010; 44, pp. 609-619. [DOI: https://dx.doi.org/10.1016/j.tra.2010.04.002]

36. Couclelis, H. From Sustainable Transportation to Sustainable Accessibility: Can We Avoid a New Tragedy of the Commons? in Information, Place, and Cyberspace: Issues in Accessibility (eds. Janelle, D. G. & Hodge, D. C.) 341–356 (Springer, Berlin, Heidelberg, 2000). https://doi.org/10.1007/978-3-662-04027-0_20.

37. Rizki, M; Basuki Joewono, T; Susilo, YO. Towards understanding travel in the digital age: A cross-dimensional one-week diary of individual virtual and physical activities in Indonesian cities. Transp. Res. Part Policy Pract.; 2024; 187, 104195. [DOI: https://dx.doi.org/10.1016/j.tra.2024.104195]

38. Farag, S; Krizek, KJ; Dijst, M. E-Shopping and its Relationship with In-store Shopping: Empirical Evidence from the Netherlands and the USA. Transp. Rev.; 2006; 26, pp. 43-61. [DOI: https://dx.doi.org/10.1080/01441640500158496]

39. Mokhtarian, PL. If telecommunication is such a good substitute for travel, why does congestion continue to get worse?. Transp. Lett.; 2009; 1, pp. 1-17. [DOI: https://dx.doi.org/10.3328/TL.2009.01.01.1-17]

40. Wang, D; Law, FYT. Impacts of Information and Communication Technologies (ICT) on time use and travel behavior: a structural equations analysis. Transportation; 2007; 34, pp. 513-527. [DOI: https://dx.doi.org/10.1007/s11116-007-9113-0]

41. Ding, Y; Lu, H. The interactions between online shopping and personal activity travel behavior: an analysis with a GPS-based activity travel diary. Transportation; 2017; 44, pp. 311-324.3727976 [DOI: https://dx.doi.org/10.1007/s11116-015-9639-5]

42. Alemi, F; Circella, G; Handy, S; Mokhtarian, P. What influences travelers to use Uber? Exploring the factors affecting the adoption of on-demand ride services in California. Travel Behav. Soc.; 2018; 13, pp. 88-104. [DOI: https://dx.doi.org/10.1016/j.tbs.2018.06.002]

43. Irawan, MZ; Belgiawan, PF; Tarigan, AKM; Wijanarko, F. To compete or not compete: exploring the relationships between motorcycle-based ride-sourcing, motorcycle taxis, and public transport in the Jakarta metropolitan area. Transportation; 2020; 47, pp. 2367-2389. [DOI: https://dx.doi.org/10.1007/s11116-019-10019-5]

44. Irawan, M. Z. & Belgiawan, P. F. Ride-hailing app use for same-day delivery services of foods and groceries during the implementation of social activity restrictions in Indonesia. Int. J. Transp. Sci. Technol. (2022) https://doi.org/10.1016/j.ijtst.2022.03.004.

45. Lee, Y. & Circella, G. Chapter Five - ICT, millennials’ lifestyles and travel choices. in Advances in Transport Policy and Planning (ed. Ben-Elia, E.) vol. 3 107–141 (Academic Press, 2019).

46. Prasetyanto, D; Rizki, M; Sunitiyoso, Y. Online Learning Participation Intention after COVID-19 Pandemic in Indonesia: Do Students Still Make Trips for Online Class?. Sustainability; 2022; 14, 1982.2022Sust..14.1982P [DOI: https://dx.doi.org/10.3390/su14041982]

47. Rizki, M., Joewono, T. B., Belgiawan, P. F. & Irawan, M. Z. The travel behaviour of ride-sourcing users, and their perception of the usefulness of ride-sourcing based on the users’ previous modes of transport: A case study in Bandung City, Indonesia. IATSS Res. (2020) https://doi.org/10.1016/j.iatssr.2020.11.005.

48. Asmussen, KE et al. An investigation of individual-level telework arrangements in the COVID-era. Transp. Res. Part Policy Pract.; 2024; 179, 103888. [DOI: https://dx.doi.org/10.1016/j.tra.2023.103888]

49. Losa Rovira, Y; Faghih Imani, A; Sivakumar, A; Pawlak, J. Do in-home and virtual activities impact out-of-home activity participation? Investigating end-user activity behaviour and time use for residential energy applications. Energy Build; 2022; 257, 111764. [DOI: https://dx.doi.org/10.1016/j.enbuild.2021.111764]

50. Beierle, F et al. Frequency and duration of daily smartphone usage in relation to personality traits. Digit. Psychol.; 2020; 1, pp. 20-28. [DOI: https://dx.doi.org/10.24989/dp.v1i1.1821]

51. Butt, S; Phillips, JG. Personality and self reported mobile phone use. Comput. Hum. Behav.; 2008; 24, pp. 346-360. [DOI: https://dx.doi.org/10.1016/j.chb.2007.01.019]

52. Wu, KD; Clark, LA. Relations between personality traits and self-reports of daily behavior. J. Res. Personal.; 2003; 37, pp. 231-256. [DOI: https://dx.doi.org/10.1016/S0092-6566(02)00539-1]

53. Couclelis, H. Rethinking Time Geography in the Information Age. Environ. Plan. Econ. Space; 2009; 41, pp. 1556-1575. [DOI: https://dx.doi.org/10.1068/a4151]

54. Motte-Baumvol, B; Belton Chevallier, L; Bonin, O. Does e-grocery shopping reduce CO2 emissions for working couples’ travel in England?. Int. J. Sustain. Transp.; 2023; 17, pp. 515-526. [DOI: https://dx.doi.org/10.1080/15568318.2022.2074326]

55. ALMEC Corporation. Jabodetabek Urban Transport Policy Integration Study. (2019).

56. Hermawati, P; Aryawan, IGMO; Sutapa, IK; Santiana, IMA. Kajian Permintaan Perjalanan Penumpang dalam Rangka Penyediaan Prasarana Sarana Transportasi Umum di Bali. J. Bali Membangun Bali; 2020; 1, pp. 179-192. [DOI: https://dx.doi.org/10.51172/jbmb.v1i3.139]

57. Soto, CJ; John, OP. Short and extra-short forms of the Big Five Inventory–2: The BFI-2-S and BFI-2-XS. J. Res. Personal.; 2017; 68, pp. 69-81. [DOI: https://dx.doi.org/10.1016/j.jrp.2017.02.004]

58. Ahmed, SK. How to choose a sampling technique and determine sample size for research: A simplified guide for researchers. Oral Oncol. Rep.; 2024; 12, 100662. [DOI: https://dx.doi.org/10.1016/j.oor.2024.100662]

59. Budlender, D. A. Critical Review of Selected Time Use Surveys. Gender and Development Program Paper No. 2. United Nations Research Institute for Social Development (2007).

60. Hirk, R; Hornik, K; Vana, L. Multivariate ordinal regression models: an analysis of corporate credit ratings. Stat. Methods Appl.; 2019; 28, pp. 507-539.4009760 [DOI: https://dx.doi.org/10.1007/s10260-018-00437-7]

61. Basu, N; Oviedo-Trespalacios, O; King, M; Kamruzzaman, L; Md. Haque, SM. The influence of the built environment on pedestrians’ perceptions of attractiveness, safety and security. Transp. Res. Part F Traffic Psychol. Behav.; 2022; 87, pp. 203-218. [DOI: https://dx.doi.org/10.1016/j.trf.2022.03.006]

62. Ma, J; Ye, X; Shi, C. Development of Multivariate Ordered Probit Model to Understand Household Vehicle Ownership Behavior in Xiaoshan District of Hangzhou, China. Sustainability; 2018; 10, 3660.2018Sust..10.3660M [DOI: https://dx.doi.org/10.3390/su10103660]

63. Piras, F; Sottile, E; Tuveri, G; Meloni, I. Could there be spillover effects between recreational and utilitarian cycling? A multivariate model. Transp. Res. Part Policy Pract.; 2021; 147, pp. 297-311. [DOI: https://dx.doi.org/10.1016/j.tra.2021.03.017]

64. Meyer, BD. Unemployment Insurance and Unemployment Spells. Econometrica; 1990; 58, pp. 757-782. [DOI: https://dx.doi.org/10.2307/2938349]

65. Singh, AC; Faghih Imani, A; Sivakumar, A; Luna Xi, Y; Miller, EJ. A joint analysis of accessibility and household trip frequencies by travel mode. Transp. Res. Part Policy Pract.; 2024; 181, 104007. [DOI: https://dx.doi.org/10.1016/j.tra.2024.104007]

66. Eurostat. E-commerce statistics for individuals. https://ec.europa.eu/eurostat/statistics-explained/index.php?title=E-commerce_statistics_for_individuals (2022).

67. Tirachini, A. Ride-hailing, travel behaviour and sustainable mobility: an international review. Transportation (2019) https://doi.org/10.1007/s11116-019-10070-2.

68. Hair, J. F., Black, W. C., Babin, B. J. & Anderson, R. E. Multivariate Data Analysis. (Pearson, Harlow, Essex, 2010).

69. Kaiser, HF. An index of factorial simplicity. Psychometrika; 1974; 39, pp. 31-36. [DOI: https://dx.doi.org/10.1007/BF02291575]

70. Chakrabarti, S. Does telecommuting promote sustainable travel and physical activity?. J. Transp. Health; 2018; 9, pp. 19-33. [DOI: https://dx.doi.org/10.1016/j.jth.2018.03.008]

71. Shabanpour, R; Golshani, N; Tayarani, M; Auld, J. Analysis of telecommuting behavior and impacts on travel demand and the environment. Transp. Res. Part Transp. Environ.; 2018; 62, pp. 563-576. [DOI: https://dx.doi.org/10.1016/j.trd.2018.04.003]

72. Schaner, S. & Das, S. Female Labor Force Participation in Asia: Indonesia Country Study. https://papers.ssrn.com/abstract=2737842 (2016).

73. Liu, C; Susilo, YO; Dharmowijoyo, DBE. Investigating intra-household interactions between individuals’ time and space constraints. J. Transp. Geogr.; 2018; 73, pp. 108-119. [DOI: https://dx.doi.org/10.1016/j.jtrangeo.2018.10.015]

74. Ho, C; Mulley, C. Intra-household interactions in transport research: a review. Transp. Rev.; 2015; 35, pp. 33-55. [DOI: https://dx.doi.org/10.1080/01441647.2014.993745]

75. Hu, Y; van Wee, B; Ettema, D. Intra-household decisions and the impact of the built environment on activity-travel behavior: A review of the literature. J. Transp. Geogr.; 2023; 106, 103485. [DOI: https://dx.doi.org/10.1016/j.jtrangeo.2022.103485]

76. Suwardani, N. P., Paramartha, W. & Suasthi, I. G. A. BALE BANJAR AND ITS IMPLICATIONS ON THE EXISTENCE OF BALI SOCIOCULTURAL COMMUNITIES (2018).

77. BPS Kota Denpasar. Jumlah Banjar Menurut Jenisnya Per Kecamatan. https://denpasarkota.bps.go.id/id/statistics-table/2/NjEjMg==/jumlah-banjar-menurut-jenisnya-per-kecamatan.html (2022).

78. Joni, I. D. M. A. B., Talip, B. A., Mokhtar, S. A. & Permana, I. P. H. The Smart Concept to Prosper the Community with the Development of Local Wisdom in the Banjar Institution. in Tech Horizons:Unveiling Future Technologies (eds. Ismail, A., Zulkipli, F. N., Husin, H. S. & Öchsner, A.) 93–101 (Springer Nature Switzerland, Cham, 2024). https://doi.org/10.1007/978-3-031-63326-3_12.

79. Post, J. C. Ethnomusicology: A Contemporary Reader. (Routledge, 2013).

80. Roberts, M., Sander, F. G. & Tiwari, S. Time to ACT: Realizing Indonesia’s Urban Potential. (World Bank Publications, 2019).

81. Mokhtarian, PL; Salomon, I; Handy, SL. The Impacts of Ict on leisure Activities and Travel: A Conceptual Exploration. Transportation; 2006; 33, pp. 263-289. [DOI: https://dx.doi.org/10.1007/s11116-005-2305-6]

82. Ministry of National Secretariat. Pemerintah Terbitkan Aturan Flexible Working Arrangement bagi ASN. https://www.setneg.go.id/baca/index/pemerintah_terbitkan_aturan_flexible_working_arrangement_bagi_asn (2023).

83. Suparman, S. The great workplace debate: Remote, hybrid or office? The Weekender by The Jakarta Posthttps://weekender.thejakartapost.com/prosperity-and-progress/2024/09/09/the-great-workplace-debate-remote-hybrid-or-office.html (2024).

84. Susilo, YO; Avineri, E. The impacts of household structure on the individual stochastic travel and out-of-home activity time budgets. J. Adv. Transp.; 2014; 48, pp. 454-470. [DOI: https://dx.doi.org/10.1002/atr.1234]

85. Saffa, A. Indonesia’s Digital Vision 2045 Guides Transformation. OpenGov Asia. https://opengovasia.com/2023/12/14/indonesias-digital-vision-2045-guides-transformation/ (2023).

86. Switzerland Global Enterprise. Industry Report: ICT in Indonesia. S-GE. https://www.s-ge.com/en/publication/industry-report/20182-ict-indonesia (2023).

87. Bandung Statistics Bureau. Bandung Population 2018-2020. https://bandungkota.bps.go.id/indicator/12/32/1/jumlah-penduduk.html (2022).

88. Cianjur Bureau of Statistics. Jumlah Penduduk Menurut Kecamatan Berdasarkan Hasil SP (Jiwa), 2020-2021. https://cianjurkab.bps.go.id/indicator/12/222/1/jumlah-penduduk-menurut-kecamatan-berdasarkan-hasil-sp.html (2022).

89. Denpasar Statistics Bureau. Denpasar Population 2018-2020. https://denpasarkota.bps.go.id/indicator/12/49/1/proyeksi-penduduk-kota-denpasar.html (2021).

90. Jakarta Bureau of Statistics. Jumlah Penduduk Provinsi DKI Jakarta Menurut Kelompok Umur dan Jenis Kelamin 2019-2021. https://jakarta.bps.go.id/indicator/12/111/1/jumlah-penduduk-provinsi-dki-jakarta-menurut-kelompok-umur-dan-jenis-kelamin.html (2022).