Exploring Gender Differences through the Lens of Spatiotemporal Behavior Patterns in a Cultural Market: A Case Study of Panjiayuan Market in Beijing, China

Abstract

Markets are more than commercial places with high mobility, greatly contributing to urban vitality, social inclusion, and even local identity, which have all been studied extensively. However, as one of the market types, the cultural market contains a highly gendered feature in our contemporary cities that has rarely been explored. Therefore, this article presents a survey that uses spatiotemporal behavior mapping (STBM) to collect gender-related behavior patterns across four sites, five activity layers, four age groups, and three time dimensions in Beijing Panjiayuan cultural market, China. Our results show that all sites are generally dominated by males. Although mobility is higher on weekends than weekdays, the number of women decreases and the number of men increases at the weekend, resulting in a higher gender ratio. The gender ratio in the 19–36 and 54+ year groups synchronized decline with mobility, whereas other age groups did not appear to be significantly affected. More specifically, in the buying and common activities layers, the ratio of men to women at the two sites with higher mobility increased, and vice versa. It is the first study to present the effect of spatial mobility on gender differences in a cultural market by using objective spatiotemporal data. Our findings provide a scientific basis for the optimization and design practice of culture-related spaces to create a more equal and inclusive urban environment.

1. Introduction

Spatial equity is a dimension of spatial justice, embodied in fairness of accessibility and participation in social resources, services, and life [1]. As a set of contrary results, equity and inequity of the space arise during the procedure of social production; that is, the root of spatial inequality is constrained within a social structure. In this situation, vulnerable groups defined by race, occupation, gender, and income are confined and excluded due to social injustices consisting of exploitation, marginalization, powerlessness, and violence. This may lead to the reality that the basic rights of partial residents for physical public spaces could not be pledged. Soja [1] emphasizes that the right to physical urban space access is a fundamental guarantee for the equal realization of the right to the city. While spatial equality plays an important role in both individual and social development, gender-specific inequalities in individual mobility, including access to and use of urban space, persist and remain [2,3]. They are also reflected in a large number of barriers based on physicality, and property, and even emotion [3]. The urban spatial inequalities that result from sociocultural constructs in particular are hidden and often invisible in the gender dimension [4].

Space and place are not neutral but gendered in varying degrees [5]. According to Valentine [6], the “geography of women’s fear” gives rise to dramatic gender differences in the perception of safety and leads to different ways of using of public space. Many studies have confirmed this. For example, women choose to travel in pairs rather than alone, reduce the amount of time spent on the street, shop at relatively upmarket establishments, or/and avoid lower/middle-class men [7,8,9]. Women have adopted a number of behavioral responses in order to navigate urban spaces more safely [10]. They adjust elements such as the extent, frequency, time, duration, and manner of presence so that a range of potential heterosexual violence, harassment, or simply “male gaze” can be avoided [7,11].

The perception of safety is not specific to women, but rather different between men and women and highly influenced by subjective spatial experiences [12,13]. Therefore, gender-based interventions in the built environment are also necessary. A study by Hernandez et al. [12] on perceived safety in the built environment of urban public spaces (London) showed that the impact of different interventions differed by gender. Of the three interventions—public toilets, solid wall removal, and graffiti—solid wall removal had a much stronger impact on women than men. Another study by Agustí et al. [13] in Spain also verified that the spaces that make men and women feel pressure or fear are different. For men, open spaces tend to cause higher levels of stress; for women, narrow streets and the presence of other people produce the greatest stress. Therefore, interventions in gender-sensitive built environments need to be implemented based on an understanding of gender differences.

The gendering of culturally relevant places is more pronounced and is influenced by different cultures in different places. Massey [5] refers to this with the example of the football field in Manchester, UK, where she suddenly realized was a place she had never ventured or attempted to venture into, a place for teenage boys. Another oft-cited example is the art gallery, filled with nude female works of the male gaze [5]. Department stores are another representative site of gendering. The array of goods attracts the gaze of women, who are transformed from the object of the general condition of being gazed at into the subject. But this power is very limited, and in the larger picture, the original department stores were an extension of the household or housekeeping. Contemporary department stores and shopping malls seem to provide places for women to participate in the public space of the city, but they are filled with the regulation and control of women [14]. The Tehran Bazaar is a male-dominated space, and women have to solidify their presence by moving deeper into the marketplace and other tactics [15].

The above studies show that culturally relevant places are more gendered and have different patterns of behavioral expression. Blindly renovating the built environment without taking full account of the gendered differences constructed by local socio-culture will not enhance the use of space for women to the greatest extent possible [15].

Apart from the perception of safety, mobility is another factor that influences gender differences in access to and use of public spaces [10]. There are two types of mobility, one of which is individual mobility, or person-based mobility, which is expressed as the movement of individuals in their daily lives. Data are generally obtained and studied using personal diaries, call records, or GPS locations [16,17]. Individual mobility is not gender neutral [18]. Insecurity and fear of being hurt are the main reasons that limit women’s travel. In addition to this, women’s transport choices are often influenced by their low income levels [10]. Gender role restrictions that lead to more domestic activity-oriented travel for women are also one of the main reasons for mobility inequality [3,19].

Another one is spatial mobility, i.e., location-based mobility. Previous studies have often defaulted to the idea that space is stable and unchanging, ignoring the fact that space is constantly undergoing a dynamic production process [20,21]. Spatial mobility implies a change in location or state, and the process of mobility is also a process of reconstructing spatial relations [22]. High mobility can lead to greater dynamism in urban economic and social interaction, but it can also mean greater uncertainty, which can lead to problems such as crimes, disorder, or social conflicts [23,24]. In social practices that operate using mobility, the complexity and uncertainty of spatial systems are significantly enhanced by the combination of social and geographical dimensions [22,25], which may exacerbate gender inequalities in space occupancy and use. For instance, a study in Brazil shows that women are more likely to encounter violence than men at the central metro station [18]. However, spatial mobility has been less considered in previous studies on gender differences due to its complexity and difficulty in measuring calculations.

China was considered to have greater gender equality in both the economic (equal work, equal pay) and political (the equal right to vote) spheres than many Western countries—at least in urban society [26]. However, some studies have looked at the range of inequalities that result from the division of roles in the family. The traditional gender division of labor, whereby men are responsible for earning money and women are responsible for running the household, has never fundamentally been challenged [26]. This has led to a double pressure on women in terms of work and domestic responsibilities, under which they not only have to sacrifice their career opportunities (including opportunities for promotion and more complex work), but also have less discretionary time. A series of studies confirmed this by revealing the gender differences in time use and daily activities [17,27]. Other studies have explored gender differences in urban consumption spaces using questionnaires or web text analysis and the findings are consistent with the above studies, namely that gender differences are mainly reflected in the fact that women have lower consumer power and less leisure time [14,28].

In general, empirical research on gender differences in the use of public space in China is limited, and apart from the abovementioned studies, most of the relevant research is still at the level of introduction to Western theories [28,29,30], and there is a lack of local-focused and place-based research. However, since gender is a product of culture rather than nature [31,32], women’s experience and use of space may differ between China and the West, metropolitan and small cities, and specific areas within cities. In addition, as mentioned above, culture-related spaces are particularly gendered, for instance, the football playground in the UK or the bazaar in Iran, and have different expressions in different cultural contexts. In this study, although the gender ratio in Beijing is nearly equal, men are still clearly dominant in this cultural market. Therefore, research on spatial gender differences based on specific local cultural and social contexts is necessary.

Over the past decades, feminist scholars have provided significant evidence supporting the claim that “how people use public space (where, when, with whom, how to get there (individual mobility/time accessibility)) is significantly gendered” [5,6,13,21]. Generally, several empirical studies conducted in both developed and developing countries have found that women tend to travel in pairs, spend less time in leisure relative space, avoid narrow streets, and stay away from environments lacking lighting more than men. The previous literature has examined the complex relationship between gender and public space use, with a particular focus on gender differences in perceptions of safety. Perceptions of safety are influenced by the environment, and spatial mobility is an important but underexplored aspect of the environment.

This work studies gender differences from a spatiotemporal and mobile space perspective in the Panjiayuan cultural market, Beijing, the capital of China, with almost 21.9 million inhabitants; 51.14% of them men and 48.86% them women [33]. The large population poses many challenges to urban planners and policymakers, particularly relating to the regeneration of existing public space to provide inclusivity and equality needs for its inhabitants.

Based on the above literature, three main hypotheses are proposed: in the cultural market,

• Sites/spaces with high mobility have low female presence;
• The female proportion decreases when mobility increases at the same site;
• The effect of mobility on gender behaviors is influenced by other environmental factors within the site (e.g., openness and light) and age groups.

This study aims to examine gender differences in activities, age groups, and time periods in cultural market spaces under different spatial mobilities. To this end, it first assesses the relationship between spatial mobility and gender activities in four sites, five activity layers, four age groups, and three time dimensions. Second, it explores spatial patterns of gender behaviors in high spatial mobility areas (i.e., Site 2 and Site 4). Understanding different behavioral patterns is the basis for design interventions, and this approach can be used in similar cultural urban spaces. This could help in identifying gender inequality public spaces with a view to improving them and thereby achieving more equitable and inclusive cities.

2. Research Methods

2.1. Site Selection

The site, Panjiayuan Market, Beijing, China (Figure 1), an arts, crafts, and antique market, was chosen from several markets (for example, the Longfu Temple area, Liulichang Antique Street, and the Daliushu Market) for the following reasons:

• The Panjiayuan market opens at a regular time, which guarantees the possibility of carrying out timed and fixed data collection;
• Over 30 years of history with the absence of excessive upscaling or excessive designer intervention resulted in the preservation of the specific social and local character of the site;
• Great mobility, both in terms of pedestrian movement and the ability to switch between tasks. Despite the lingering effects of the COVID-19 pandemic and the lockout policy, it manages to maintain a high level of pedestrian movement. (With an average of 12,000 people/day on weekdays, and an average of 26,000 people/day on weekends, as of July 2021);
• High heterogeneity in terms of both spaces and social groups;
• The variety of functions, for example, shopping, strolling, and sightseeing.

The site is located on the west side of the East Third Ring Road, Chaoyang District, which is the second strongest economic district in Beijing, with a gross regional product of 7 million in 2020 (Beijing Chaoyang District Statistical Yearbook 2021). The good economy contributes to the concentration of population and prosperity of consumption, especially with regard to cultural commodities, which jointly facilitate the mobility of the site. A high degree of heterogeneity among social groups and a wide variety of functions are associated with the location. Panjiayuan market is bordered by a variety of residential neighborhoods, including five within a 500 m radius and 16 within 15 min walking distance, as well as amenities including metro stations, hospitals, and several dining options.

The following are the principles underpinning the choice of fixed spots for spatiotemporal behavior mapping (STBM) in this study:

• High spatial temporal mobility, i.e., a high flow of people or a high amount of variation; high spatial vitality, where activities continue to be frequent during the observation period or flourish regularly at a certain time; and high heterogeneity of behavioral activities.
• There are differences in spatial and temporal characteristics between different observation sites.

Based on the above principles, four typical ground stall areas with the most active behavioral activities in the market were selected for data collection in this study (Figure 2).

2.2. Design and Methodology

The methodology of STBM draws on the study of street vending and walking behavior in Yuncheng, China [34]. STBM is a method of studying and observing behavior from a temporal and spatial perspective onsite. Observation and behavior mapping have been widely used to learn how space characters support [35] or impact user behavior in a specific setting rather than generally across a site. In addition, it provides a visual and statistical representation of behavioral patterns [36]. The basic behavior mapping approach is a four-step process:

• Make an accurate scale base map of the observation area.
• Define the types of activities.
• Set the rules for the coding system.
• Set a repeated observation schedule for a specific time.

The above methods have been used in many observational studies. However, due to the high mobility and heterogeneity of the market and the simultaneous observation and data collection from four sites, there are several problems regarding the application of the above method:

• The coding system is too complex.
• Difficult to unify the judgment and determination of activities.
• The limited speed of manual tagging results in easy loss of behavioral activity data.
• After manual marking, it cannot be checked at a later stage, and data quality is difficult to guarantee.

Based on the above reasons, the manual records were replaced with a photographic model and entered into the GIS system at a later stage to reduce errors caused by instant recording. This initiative also guarantees the checkability of data. The STBM has devel-oped from the above process. In addition, this study explores the analysis of the data col-lected by the STBM from a gender perspective and adds an objective analysis of spatial mobility (see Section 2.5.1) to the subjective analysis of space in the original methodology. Furthermore, a summer workshop was organized, and several students were recruited to ensure that data collection could take place simultaneously at each of the four sites.

The overall principles of ethics and privacy are considered twofold: on the one hand, permission to film was obtained in advance from the market’s management, who issued permits to all filming crews present. At the same time, a filming statement was posted in a prominent place near the filming location, explaining how and for what purpose the filming would take place. On the other hand, the data collection process minimized disruptions to onsite behaviors as much as possible.

2.3. Data Collection

The data collection consists of two parts: the pilot study and the official collection. The official collection lasted for 1 week (from 13 July 2021 to 18 July 2021) to cover the time that daily activities take place. The collection mainly consisted of the following four aspects:

• Pilot study: A pilot study was conducted 2 days before the official collection (10–11 July 2021). The location, height, and angle were changed a number of times before being settled upon for the pilot project. This is due to the necessity for maintaining shot clarity (e.g., minimizing umbrella shade) and minimizing interruption of the venue’s activity in during busy times.
• Base map: to clarify the interconnection between behavior and a specific setting, the base map marks also included immovable environments (such as buildings or structures, ground coverings, or ground cracks,) and facilities that remained in the same position continuously during the observation period (such as sunshade holes, seats, litter bins, and other objects that have an obvious impact on activities). Due to the advantages of geographic information system (GIS) technology, drawing could be conducted along with data collection at the same time.
• Data were collected at four time periods every day, each lasting one hour. The specific time periods were determined based on the daily rhythm of market activity: after the market opens (10:00–11:00 a.m.), lunch hours for most vendors (12:00–1:00 p.m.), peak pedestrian traffic (4:00–5:00 p.m.), and before the market closes (7:00–8:00 p.m.). In each one-hour period, one photo was taken every 5 min, for a total of 13 photos per hour.
• Because the ground at Panjiayuan is flat, it was difficult to find a high spot for taking photos and observing behaviors. Moreover, all observation places were covered with shade cloths, sunshades, or canopies. Therefore, four tripods were employed, raising each camera and maintaining a constant height throughout the course of seven days in order to increase the database’s accuracy.
• Other information used to assist in understanding the relationship between the activity and the environment was recorded using subjective evaluation sheets and notes. For example, during the rain showers on 13 July, Site 1 had a sharp drop in footfall due to water on the ground and the lack of a rain barrier, while Site 2 had an increase in footfall due to the presence of a canopy to protect it from the rain.

2.4. Database Set

Using the coding system, behavioral information and information about the activities’ attributes were coded.

2.4.1. Behavior Data

Behaviors were categorized into five layers from a total of 32 activities. In the coding and entry process, these five tiers of activities have been categorized in an inverted pyramid according to ease of identification (Figure 3). The selling layer is the easiest to identify, i.e., vendors selling various items; buyers are the customers; special activities (i.e., kids playing, walking with stroller/child, or live streaming) are usually related to children, strollers, or professional live streaming equipment; special people actually refers to a specific activity brought about by a special group of people, usually identified by their uniform—for example, market surveillance refers to the supervision of the market brought about by market surveillance, and cleaning refers to the cleaning activity that accompanies cleaning—thus, the term “special people” as a code word in action still refers to activities such as those mentioned above; the remaining population is categorized by common activities, including walking, sitting, squatting, standing, and cycling.

2.4.2. Participant Characteristics

This study estimated people’s gender (male/female) and age (four groups): adolescents (0–18 years), young adults (19–36 years), adults (37–54 years), and older people (54+ years). As the observations were made from a distance, it was impossible to determine people’s age with any precision, so the age groups were divided across a wide range. The division into specific age groups follows the results of previous studies on different age groups’ walking behavior in China [37]. Early in the study, the “working age population” was divided into three groups: late adolescents (aged 18–25), young adults (aged 26–35), and middle-aged adults (aged 36–59). We merged the first two age groups and set the final age group according to the current average retirement age in China, which is 55 years old.

Judgments about age and gender were largely based on appearance, manner, and clothing, which can vary considerably among different groups. Therefore, some misclassifications may occurred for these two variables.

2.5. Data Analysis

2.5.1. Statistical Analysis

To explore the relationship between spatial mobility and gender, this study explores the effect of mobility on the proportion of gender in five behavioral layers across four observation sites and four age groups, under three time dimensions.

The mobility of space is the change of mobile elements in space per unit of time. In this paper, for the measurement of spatial mobility, we refer to the methodology of Yonglin et al. [38] for the study of the mobility of individual cities. They argued that the mobility of an individual city cannot be measured, further using transport flows, shift counts, etc., because it takes place between two cities, from one to another. The mobility of a city is related to peoples tendency to move. Thus, kinetic and potential energy are interconverted using a physics formula. The height difference is also analogized to the increment of each indicator in the city.

$$\mathrm{V}=\sqrt{\Delta Q}$$

where V represents the mobility, and $\Delta Q$ refers to the amount of change in the indicator over a given period.

To determine how mobile the Panjiayuan market space is, two types of flows—population flow and activity flow—as well as two observations—population number and activity number—were chosen. ΔQp represents the amount of change in the number of people and $\Delta Qa$ represents the amount of change in activity. V is mobility.

Population flow and activity flow were selected as the variables used for measuring space mobility. Specifically, the mobility of a particular period is the square root of half the sum of the population’s activity that we recorded. The arithmetic formula is shown below. $\Delta Qp$ represents the change in population, and $\Delta Qa$ represents the change in activities.

$$\mathrm{V}=\sqrt{\frac{1}{2}\ast (\Delta Qp+\Delta Qa)}$$

2.5.2. Spatial Analysis

For the spatial analysis, we chose Sites 2 and 4, which had relatively high mobility, to explore the behavioral and spatial distribution patterns of different genders in a high-mobility environment.

3. Results

A total of 8587 marker points were recorded for all sites (

Table 1. Gender, age, and activity.

	Categories	Number of Points	Proportion
Gender	Male	6214	72.37%
	Female	2373	27.63%
Age	0–18	235	2.74%
	19–36	1999	23.28%
	37–54	4778	55.64%
	54+	1575	18.34%
Activity	Selling	3843	44.75%
	Buying	1625	18.92%
	Special activities	96	1.12%
	Special people	51	0.59%
	Common activities	2972	34.61%

). Of these, Site 1 had 1774; Site 2 had 2176; Site 3 had 1170; and Site 4 had 4424. The dominant gender was male (72.37%), while the female portion (27.63%) was nearly one-third of that. The majority of respondents belonged to the 37–54-year-old age group, while only 2.74 percent of the respondents belonged to the 0–18-year-old age group.

In terms of activities, selling and common activities were the majority of activities (44.75% and 34.61%). Buying, special activities, and special people accounted for 18.92%, 1.12%, and 0.59%, respectively.

This section is divided using subheadings. It provides a concise and precise description of the experimental results, their interpretation, as well as the experimental conclusions that can be drawn.

3.1. Time, Mobility, and Gender Ratio

This subsection develops a study of differences in the proportion of gender presence across different time dimensions and different activity layers. Specifically, it analyzes the relationship between mobility and the gender ratio within the venue from three different time dimensions. A full week (6 days), weekends and weekdays, and four time periods during the day constitute the three time dimensions. This separate analysis is deemed necessary, as the Panjiayuan Market is typically a cyclical market with a large difference in mobility between weekends and weekdays, as well as different periods of the day.

3.1.1. Six Days

According to

Table 2. Mobility and gender ratio in activities across four sites on six days.

		Site 1	Site 2	Site 3	Site 4
Mobility		29.95	33.07	24.37	41.77
Gender ratio (male/female)	Selling	1.58	2.75	4.17	2.83
	Buying	3.39	3.79	3.47	3.66
	Special activities	0.50	3.00	2.20	1.63
	Special people	-	-	-	2.10
	Common activities	1.99	2.57	2.40	2.46
	Total	1.88	3.02	3.21	2.70

, on the six days, Site 1 and Site 3 have lower mobility, and Site 2 and Site 4 have higher mobility. In overall gender ratio, Site 1 and Site 4 are lower than Site 2 and Site 3. That is, the overall gender ratio and mobility do not match our hypothesis. In addition, it is noteworthy that the gender ratio patterns are accordant with mobility in the buying and common activities layers, indicating that mobility has a different effect on the presence of different genders. The percentage of men who make purchases and engage in common activities increases when mobility is high. Selling and special activities, however, do not show an effect on mobility.

3.1.2. Weekdays and Weekends

As can be seen from

Table 3. Mobility and gender ratio in activities across four sites on weekdays and weekends (abbreviations: WEND—weekends; WKD—weekdays).

		Site 1		Site 2		Site 3		Site 4
		WKD/4	WEND/2	WKD/4	WEND/2	WKD/4	WEND/2	WKD/4	WEND/2
Mobility		11.71	13.87	13.75	13.38	9.09	12.26	17.65	16.99
Gender	Selling	0.76	4.67	2.55	3.78	3.00	6.06	2.54	3.90
ratio	Buying	2.70	4.77	3.59	4.35	2.11	3.49	2.20	5.30
(male/female)	Common activities	1.78	2.97	2.36	2.93	2.13	2.85	2.02	2.35
	Total	1.18	4.10	2.78	3.70	2.60	4.12	2.31	2.93

at Site 1 and Site 3, mobility increased on weekends compared to a weekdays, whereas at Site 2 and Site 4, mobility slightly decreased. The ratio of men to women rose to varying degrees at all four venues over the weekend, with Sites 1 and 3 witnessing greater mobility and a greater increase in ratio of men to women. Moreover, the increase in gender ratio varies for different activities. The gender ratio in the selling and buying layers increases more sharply than common behaviors on weekends.

The increase in gender ratio is not only because of the rise in the number of males but also a result of the decrease in female numbers (

Table 4. The average number of males and females across four sites on weekdays and weekends (abbreviations: WEND—weekends; WKD—weekdays).

Site 1			Site 2		Site 3		Site 4
	WKD/4	WEND/2	WKD/4	WEND/2	WKD/4	WEND/2	WKD/4	WEND/2
Male	139.50	299.50	271.50	275.50	107.75	230.50	420.75	416.50
Female	117.75	73.00	97.75	74.50	41.50	56.00	182.00	142.00

). Furthermore, as most of the people are of working age, they should have more time for recreational activities on weekends, and in addition, the proportion of reductions among women varies considerably from group to group. The female vendor group (37.63%) declined to a much greater extent than the nonvendor group (6.4%) over the weekend (Figure 4).

3.1.3. Four Time Periods

The mobility of the four measured time periods in a day is closely linked to the market’s daily rhythm. As shown in

Table 5. Mobility and gender ratio in activities across four sites at four time periods.

	Time Periods	Gender Ratio (Male/Female)	Mobility
Site 1	10–11 a.m.	2.18	16.37
	12–1 p.m.	1.87	17.32
	4–5 p.m.	1.73	16.03
	7–8 p.m.	1.54	9.49
Site 2	10–11 a.m.	3.38	17.85
	12–1 p.m.	3.58	18.48
	4–5 p.m.	3.06	18.10
	7–8 p.m.	1.41	9.59
Site 3	10–11 a.m.	3.90	13.56
	12–1 p.m.	3.33	12.75
	4–5 p.m.	2.89	12.92
	7–8 p.m.	2.58	9.97
Site 4	10–11 a.m.	3.52	21.74
	12–1 p.m.	2.72	23.51
	4–5 p.m.	2.85	23.07
	7–8 p.m.	1.43	14.68

, spatial mobility was highest at 12–1 p.m. (except Site 3) and lowest at 7–8 p.m. At around 12 to 1 o’clock in the afternoon, there is a large number of tourists, and the traders become more mobile since it is lunchtime. By 7 or 8 o’clock in the evening, most of the traders have already packed up or are in the process of packing up their goods. The highest mobility in Site 2 and Site 3 is also associated with a higher proportion of males and vice versa. This result is consistent with the hypothesis of this study.

A high ratio of males to females can be observed in Sites 1, 3, and 4 between 10 and 11 a.m., one hour after the market opens. Additionally, unlike Sites 2 and Site 3, Sites 1 and Site 2 do not experience peaks in gender ratios during their spatial mobility.

3.2. Age Groups

This subsection shows the difference in gender ratios by age group at the four sites and on weekends versus weekdays. Mobility has different influences on gender ratio in different age groups. As noted earlier, mobility in Site 2 and Site 4 decreased slightly on weekend. The gender ratio in the 19–36 year group and 54+ year group synchronized decline with mobility, while the gender ratio in the 37–54 year group and in the 0–18 year group, which make up the majority of the market, continued to rise (

Table 6. Mobility and gender ratio in different age groups (abbreviations WEND—weekends; WKD—weekdays).

		Site 1		Site 2		Site 3		Site 4
		WKD/4	WEND/2	WKD/4	WEND/2	WKD/4	WEND/2	WKD/4	WEND/2
Mobility		11.71	13.87	13.75	13.38	9.09	12.26	17.65	16.99
Gender ratio (Male/Female)	0–18	0.65	1.71	1.20	4.00	1.17	5.00	0.69	2.22
	19–36	0.99	1.36	0.95	0.50	1.51	1.42	1.95	0.89
	37–54	0.86	3.17	4.46	12.76	1.76	2.17	3.17	4.33
	54+	21.43	27.75	2.34	0.72	9.68	13.95	3.84	3.17

). This may indicate that the presence of people within functional places is influenced by other factors. Moreover, gender ratio increases in parallel with mobility in all age groups at Site 1 and Site 3 (Table 6).

3.3. Spatial Distribution

This subsection examines the gender differences in the spatial distribution of the three main behaviors (selling, buying, common activities) in Site 2 and Site 4, where mobility is high during the period of the week.

3.3.1. Same-Sex Aggregation in Selling and Buying

Regarding gender differences in buying and selling behaviors, we found same-sex aggregation in buying and selling behaviors. Consider Site 4 (Figure 2), which has the most data points recorded, the largest observation area, and the highest mobility.

As we can see in Figure 5, vendors and shoppers of the same biological sex tend to be closer to each other, clusters 3 and 5 show the male gathering situation and cluster 1 shows the female cluster. At the same time, cluster 4 shows that female vendors rarely move closer to male buyers. However, it is notable that female buyers approach the male vendor’s stall in cluster 2.

In addition to the physical environment, this finding suggests that people themselves, as elements within the environment, influence how space is used and experienced by others. Highly mobile environments, in addition to inhibiting the presence of women themselves, also differ in their restrictions on women from different identity groups. The inhibitions on female vendor activity are likely to be greater than those on female non-vendor groups (e.g., tourists, buyers, etc.).

3.3.2. Walking and Other Common Activities

Regarding common activities, we define walking as a moving behavior and sitting, standing, and squatting as a staying behavior. Cycling is also classified as a staying behavior in this sense, as there no riding takes place in the market and most of the tricycles are simply parked to facilitate the loading and unloading of goods by the vendors.

The results of the behavioral map (Figure 6) show that when the environment was narrow, cramped, and dimly lit, the number and proportion of females decreased. In Site 2, at the entrances and exits, the crowd was clustered, but the gender distribution was not characterized. In the center region, there were much fewer women. In addition, the number of people moving in the middle section far exceeded the number of people staying, both for men and women. As can be seen from Figure 6, the proportion of staying behavior was also lower in the middle section than at the entrances and exits. Moreover, the variety of staying behavior was richer for males than for females in Site 2. As shown in Figure 6, females had two types of staying behavior, and males had four.

Site 2 (as shown in Figure 2) is shaded by buildings above and to the east, has a low overall height and a feeling of oppression, and is most weakly lit in the center due to a lack of lighting facilities. The oppressiveness of the space above and the darkness of the environment are probably the main reasons for inhibition.

4. Discussion

Our hypothesis was that the sites/spaces with high mobility have a low proportion of females, and this was partly supported. The gender ratio of buying and common activities in six-day periods is statistically evident. Mobility does not appear to have a significant effect on sales, special activities, and special people’s relative activities.

The hypothesis that the female proportion decreases when mobility increases at the same site was fully supported. The mobility of Sites 1 and 3 rose over the weekend and in tandem with this, the proportion of females in all three main activities layers fell (Table 3). It is worth noting that, while the mobility of Site 2 and 4 declined over the weekend, the proportion of females in their venues also declined. This might be because the increase in gender ratio jointly results in a decreased number of females and an increased number of males, possibly due to gender immobility, which is related to social role restrictions including housework or companionship needs [2,17,27]. As a result of the nature of the items sold at the Panjiayuan market, trips made there are purely leisure activities rather than domestic activities such as in supermarket visits. In addition to the higher spatial mobility, crowd density, and heterogeneity of weekend venues, and thus gender exclusion, we speculate that this is most likely due to the following facts:

• Household and childcare responsibilities are still predominantly female [39] and women have increased tasks in terms of caring for their families during weekend breaks compared to weekdays, (e.g., children spend the day at school during the weekdays and are at home on weekends and break), thus reducing their available time and opportunities for going out;
• The family structure of women varies by age. While young women remain subordinate to their families of origin, middle-aged and elderly women are more restricted to domestic duties.

Additionally, the peak in mobility and gender ratio is not synchronized between Site 1 and Site 4, but rather is highest at Site 2 and Site 3 when mobility is lowest. Site 2 is a bookstall specializing in old books, and the site provides bookcases for vendors to store their books so that they do not need to move their goods daily. Site 3 has a single-story shop on one side and, similarly, daily cargo handling is not necessary. However, the vendors at Site 1 and Site 4 are required to bring their goods into the market each day when it opens. The usual means of delivery are suitcases and tricycles, so the mixing process may require more physical labor. As a result, the work of moving goods may be more often carried out by men. As a result, both sites have the highest proportion of males at 10–11 a.m. and peak mobility is at 12–1 p.m. This supports our previous theory that other psychic elements present on the site also influence how gender is affected by migration.

The results of the behavioral map also confirm these conjectures. As the results show, during daytime (10–11 a.m., 12–1 p.m., 4–5 p.m.), open, bright, and quiet environments had a higher proportion of women than dark, oppressive, and crowded environments, and they also had a greater tendency to stay longer. The proportion of females increased at all four sites during night periods (7–8 p.m.), regardless of whether the lighting was present or not. This result further supports previous research on feminine fear in public spaces [3] and confirms that similar feminine fears are present in our culture.

In addition, it is also important to note, in conjunction with the results of this study, that the proportion of women during night periods (7–8 p.m.) increased rather than decreased, which is likely to indicate that it is not the darkness itself or the reduced visibility that women “fear” [6], or in other words, have a low safety perception of [4]. In this case, the darkness in low mobility, open or semi-open public spaces does not seem to arouse the kind of stress or fear in females that excludes them from that space.

5. Strengths and Limitations

The major strength of this study is that, as far as we know, it is the first to investigate spatial mobility in public spaces with GIS-based behavior mapping and observation methods. It could open a path to evidence-based planning, design, and research in high mobility spaces. Nevertheless, several limitations were also recognized. First, our study was only conducted during summer and is not representative of conditions during winter. Secondly, vendors rarely move from their stalls, so data are often collected in duplicate and using both statistics and mapping.

6. Conclusions

We investigated gender-specific usage of space by gender in Panjiayuan Market using STBM. It was found that men outnumbered women at all sites and periods of the day. On weekends, when the market is significantly more mobile, the average daily number of men is higher than the average weekday number, while the number of women is lower. Among the vendors, the number of males significantly exceeded that of females on weekends, and the number of females declined significantly. Same-sex aggregation was generated between buying and selling behaviors, with buyers and sellers of the same sex being physically closer to each other and the opposite sex being relatively farther apart. Female vendors had the lowest level of self-mobility.

According to the results, in male-dominated spaces, such as the Panjiayuan market, increased mobility may further restrict women’s mobility and accessibility, which exacerbates spatial inequalities. In addition, this limitation remains stronger for females from relatively disadvantaged groups, such as the elderly or those with lower socioeconomic status. Based on these results, it is believed that more storage space, more convenient handling tools, or other more physically disadvantaged-friendly policies for handling goods in the market could help to increase employment opportunities for women in this space. In addition, more lighting in dimly lit spaces with high mobility also seems to be a possible measure for reducing female fear. Finally, more open spaces not only increase the proportion of women present but also make them more inclined to stay.

To create a more inclusive and equitable society, this study, therefore, advises that, when focusing on gender gaps in highly mobile public spaces, planners, landscape architects, urban designers, and social and public health researchers should be mindful of the specific needs of particular places and the people who use them, especially vulnerable groups. No generalizations can be made about the transformation of spaces, gender-sensitive design should be location-based and culture-based.

For future research, as the STBM methodology only collects objective data, the next step could explore subjective perceptions of market participants in conjunction with interviews. Specifically, the perceptions of each gender and the reasons for them. Furthermore, classification regarding age groups could be made more accurate using sampling. In addition, this paper is our first exploration of gender and behavior in cultural markets; the intention behind spatial analysis using mapping is the exploration of new ideas and novel findings. Further research could adopt spatial autocorrelation approaches to explore clustering across time, age, or other activities in cultural markets.