The Time Use Survey of Statistics Korea (KOSTAT) collected the TAP information of 32,000, 21,000, 27,000, and 29,000 subjects from 17 provinces and metropolitan cities in Korea in 2004, 2009, 2014, and 2019, respectively [10]. The study population were selected from 800 to 850 areas of Korea, and stratified classification was used to choose fifteen households in each area. Households were selected via systematically extracting points at regular intervals within each survey area, with the regional and housing type characteristics in consideration. For each survey area, stratification criteria were established based on gender and age groups, and the sample was allocated to reflect the actual composition ratio. In each household, residents above 10 years old who agreed to participate were recruited. The TAP information of 4036, 2610, 3337, and 2793 subjects collected from Seoul in 2004, 2009, 2014, and 2019, respectively, was selected for this study.

The survey questionnaire consisted of nine main categories of sociodemographic questions: (i) personal care management; (ii) occupation; (iii) education level; (iv) household chore activity; (v) family caregiving; (vi) volunteer activity or internship; (vii) social activity; (viii) leisure activity; and (ix) transportation. The nine main categories were subdivided into 50 subcategories, which were further categorized into >120 items (Table S1). The time-location was classified by the following microenvironments: (1) time spent at residential indoor, other indoors, using transportation, and outdoors in 2004 and 2009; (2) time spent at residential indoor, workplace or school, restaurant or bar, other indoors, using transportation (bus, subway, train, taxi, private vehicle, and etc.), and outdoors in 2014 and 2019.

The data was collected using a face-to-face survey method via trained researchers, which involved two steps: (i) interview surveys and (ii) time-diary methods [10, 17]. Participants aged 10 and above recorded their activities and time-location information in a diary every 10 min for 48 h period (Fig. S1). Before initiating the surveys, respondents were instructed by the researchers to ensure that no time-diary entries were to be omitted. Researchers immediately reviewed the collected survey forms on-site for any inconsistencies, missing responses, and errors. Any discrepancies or omissions were rectified via follow-up visits or phone calls. Any part of the survey form requiring correction was marked for review and processed according to the protocol. After completing the reviews, the researchers assigned activity and location classification codes on the data provided by the respondents.

A total of 25,552 person-days diary data of the Seoul population were collected during weekdays and weekends in 2004, 2009, 2014, and 2019, respectively. In 2004, 8072 person-days data were collected during September on weekdays (4848 person-days) and weekends (3224 person-days). In 2009, 3218 and 2002 person-days data were collected during March and September, respectively, on weekdays (3113 person-days) and weekends (2107 person-days). In 2014, 1628, 3204, and 1842 person-days data were collected during July, September, and December, respectively, on weekdays (3984 person-days) and weekends (2690 person-days). In 2019, 1572, 2252, and 1762 person-days data were collected during July, September, and December, respectively, on weekdays (3388 person-days) and weekends (2198 person-days). For comparison, person-day data collected in September were extracted for this study.

The TAP information of 4401 subjects were collected from Seoul during 2022 for the Korean Air Pollutant Exposure model (KAPEX) project. Survey districts were selected using the enumeration data from KOSTAT. Households were selected via systematically extracting points at regular intervals within each survey area, with the regional and housing type characteristics in consideration. For each survey area, stratification criteria were established based on gender and age groups and the sample was allocated to reflect the actual composition ratio. In each household, residents above 10 years old who agreed to participate were recruited.

The survey questionnaire consisted of demographic characteristics (gender, age, etc.), personal characteristics (height, weight, medical issues, etc.), and socioeconomic characteristics (occupation, housing type, income, educational level, etc.). Time-location was classified by the following microenvironments: (i) residential indoor; (ii) workplace or school; (iii) kindergarten; (iv) restaurant; (v) other eating places (café and bar); (vi) study places (private educational institute, library, study café, etc.); (vii) shopping places (shopping mall, market, department store); (viii) leisure places (PC room, billiard club, screen golf club, etc.); (ix) other places; (x) outdoor; and (xi) transportation (bus, subway, train, taxi, private vehicle, etc.).

Following the steps established by KOSTAT, data was collected via trained researchers using a face-to-face survey method that involved (i) interview surveys and (ii) time-diary methods. In addition, mobile and online platforms were utilized to assist in recording the 24 h activities. Participants aged 10 and above recorded their activities and time-location information in a diary every 10 min for 48 h period. In accordance to the protocol established in KOSTAT, researchers instructed participants on filling out time-diaries, and promptly addressed any errors or inconsistencies in the completed surveys through immediate checks and follow-up actions.

Diary data for 8766 person-days of the Seoul population were collected: 2194, 2202, 2172, and 2198 person-days were collected during March, July, September, and December in 2022, respectively, on weekdays (4403 person-days) and weekends (4363 person-days). For comparison, person-day data collected in September were extracted for this study.

The microenvironment information collected in 2004, 2009, 2014, 2019, and 2022 is summarized and reclassified in Table 1. The microenvironmental classification of the KOSTAT survey in 2004 and 2009 included information of residential indoors, other indoors, transportations, and outdoors; since further subdivision was unlikely, microenvironmental data from 2014, 2019, and 2022 were reclassified into four microenvironments. The microenvironments investigated in 2014 and 2019 were reclassified as follows: (i) residential indoor included occupation in residence; (ii) workplace or school, restaurant or bar, and other indoors were classified as other indoors; (iii) using bus, subway, train, taxi, private vehicle, etc. was classified as using transportation; and (iv) outdoors included all outdoor places. Microenvironments investigated in 2022 were reclassified as follows: (i) residential indoor included occupation in residence; (ii) workplace or school, kindergarten, restaurant, other eating places (café and bar), study places (private educational institute, library, study café, and etc.), shopping places (shopping mall, market, department store), leisure places (PC room, billiard club, screen golf club, and etc.), and other places were classified as other indoors; (iii) using bus, subway, train, taxi, private vehicle, and etc. was classified as using transportation; and (iv) outdoors included all outdoor places.

Across the years, data collection seasons varied (September in 2004; March and September in 2009; March, September, and December in 2014 and 2019; March, July, September, and December in 2022), with only September being consistently represented. Despite the notable seasonal variations in TAP characteristics, due to limitations in cross-seasonal comparisons and to ensure comparability, September data collected in different years were selected for comparison in this study. To check the normality of distribution, the Kolmogorov-Smirnov test was conducted on the data collected in 2004, 2009, 2014, 2019, and 2022. As all tests showed statistically normal distributions (p < 0.001), arithmetic mean (AM) and standard deviation (SD) were used. Analysis of variance (ANOVA) was employed to determine the difference in the amount of time spent in each microenvironment in different years. For the post-hoc test, Scheffé’s method was used to evaluate the statistical difference between the microenvironmental time spent in each year. All statistical analyses were conducted using Rex-software version 3.3.1.1 (Rexsoft, Co. Ltd., Seoul, KR). A p-value of <0.05 was considered statistically significant.