What Makes a Pocket Park Thrive? Efficiency of Pocket Park Usage in Main Urban Area of Nanjing, China

Abstract

Pocket parks, recognized globally as compact yet multifunctional green spaces, promise a range of urban benefits. To realize these effectively, planners must understand the factors that drive park usage. However, development priorities vary across regions, necessitating analysis tailored to specific contexts. Existing research on park usage predominantly focuses on factors either external (factors outside the park’s boundaries, such as location and surrounding urban fabric) or internal (factors within the park’s boundaries, pertaining to design, amenities, and management), leaving room for refinement in indicator selection and model construction. To address this, we developed a comprehensive analytical framework incorporating 22 macro-, meso-, and micro-level factors spanning internal and external influences. This study investigated 40 pocket parks in Nanjing’s main urban area, employing visitor frequency and activity type diversity as quantitative indicators of usage efficiency. Park usage efficiency was compared for weekdays and weekends. Using correlation and regression models, we examined primary factors including accessibility, surrounding environment, layout, landscape features, amenities, and maintenance. The results showed that weekday and weekend patterns vary significantly in terms of visitor frequency and activity type diversity. The key determinants of pocket park usage efficiency were identified: proportion of recreational space (r = 0.609 on weekdays, r = 0.573 on weekends), plant species richness (r = 0.699 on weekdays, r = 0.761 on weekends), seat facility density (r = 0.645 on weekdays, r = 0.654 on weekends), and maintenance quality (r = 0.630 on weekdays, r = 0.667 on weekends). Service area coverage, green space ratio, and garbage bin density showed weaker but significant correlations. Based on these findings, targeted strategies were proposed to better accommodate diverse regional land-use demands.

1. Introduction

Pocket parks are targeted micro-interventions designed to revitalize underutilized or neglected urban spaces [1]. They embody a highly practical strategy for enhancing land-use efficiency, particularly within densely built environments, by transforming small, residual parcels into valuable public assets [2]. Prioritizing human-centered design principles, these compact green spaces offer accessible spots for entertainment and social interaction [3]. They bring nature and community benefits to neighborhoods, balancing the impact of large urban developments [4]. Consequently, pocket parks continue to serve as inspirational models, motivating cities worldwide to integrate and cultivate adaptable, human-scale green ecosystems into their urban fabric [5].

The European Green Deal [6] has positioned urban regeneration and land optimization as dual pillars for achieving climate neutrality under the European Climate Law. Integral to the EU’s sustainable transition program, this legislative framework mandates systemic urban carbon footprint reduction while preserving ecological capacity. Within this paradigm, pocket parks act as community-level green spaces that apply circular urbanism principles locally. They can mitigate heat islands through targeted greening, strengthen biodiversity corridors, and accelerate decarbonization at the neighborhood level.

While urban pocket parks have become established as essential infrastructure, opportunities remain to enhance environmental quality. Current challenges include declining spatial layout effectiveness, fragmented green network connectivity, and inefficient space usage that wastes resources [5]. Prevailing design approaches often prioritize aesthetic appeal over functionality, relying on subjective decisions rather than evidence-based methods. This highlights the need to systematically evaluate operational standards across representative parks.

Analyzing pocket park usage patterns and their environmental drivers enables the development of targeted optimization strategies to improve functional performance in urban green space management. Existing research defines the efficiency of pocket park usage through two key aspects: usage intensity and usage flexibility. Usage intensity refers to how many people visit the park within a specific time and space, and can be used to calculate a park’s capacity. Usage flexibility refers to how well the park supports diverse activities, reflecting balanced and adaptable resource use [7,8]. Empirical studies have revealed how usage varies by time, location, and types of activity; also variable are the factors influencing these patterns, and why they matter for urban planning and public health [8,9,10,11,12,13,14,15,16,17,18]. The question of how to integrate these diverse factors into a unified, multi-level analytical framework to quantify systematically their combined impact on usage efficiency metrics remains significantly under-researched.

While prior research has predominantly examined external or internal factors affecting pocket parks in isolation, this study establishes a comprehensive conceptual framework for usage efficiency and develops an influencing factor model to quantify relationships between different built environment characteristics and efficiency metrics. Using correlation and regression analyses, we investigate how macro-, meso-, and micro-level factors affect pocket park usage patterns. The study focuses on two questions: (1) How do usage efficiency characteristics differ between weekdays and weekends? (2) What factors affect pocket park usage efficiency? By quantitatively validating a multi-level framework of built environment features, this research provides evidence-based strategies for optimizing pocket park design and management.

This study follows a standard empirical research structure. The introduction identifies research gaps and objectives regarding pocket park usage efficiency. The literature review synthesizes the state of existing research around definitions, efficiency dimensions, and influencing factors. The methodology details the sample selection, data collection procedures, and model construction. The Section 4 presents findings on usage frequency, diversity, and spatiotemporal variations, and identifies key influencing factors through correlation and regression analyses. The discussion interprets the results and proposes optimization strategies. The conclusion summarizes major findings, acknowledges limitations, and suggests directions for future research.

2. Literature Review

2.1. Definition of Pocket Parks

Research on small urban green spaces has been conducted internationally for a long time, with different countries proposing diverse concepts of pocket parks based on respective urban development needs and cultural backgrounds. The concept of “pocket parks” emerged in the United States in the 1960s, referring to small open spaces around residential areas or commercial buildings [19]. Examples include Paley Park in New York [20]. American pocket parks emphasize intensive design to meet the needs of community outdoor activities, children’s recreation, and social interaction [19]. Both the UK and European countries tend to utilize fragmented urban spaces flexibly, viewing pocket parks as small public spaces under 0.4 ha that are accessible to everyone and provide places for people to escape from urban noise [21]. Examples include the Bryghusgrunden Pocket Park in Copenhagen [22], the Portiekje Stads Park in the Netherlands [21], and the Falata/Prusa Street Park and “Butterfly Garden” in Kraków [1]. These all focus on combining ecological conservation with daily recreational functions for citizens. Japan proposed the concept of “neighborhood parks” after World War II, referring to the vibrant core of a community, centrally located to serve residents living within 600 m. Governed by legal standards (typically 2 ha), accessibility is assessed on the principles of space syntax integrating schools, commerce, and public services [23]. Examples include the Sumida River Mini-Park in Tokyo [2], which serves communities within a 250 m radius and transforms alleys into accessible linear spaces through frequent use. Despite the terminological differences, these concepts all focus on the daily service function of micro-green spaces, addressing the issue of insufficient public space in high-density cities through flexible scale, high-density distribution, and multifunctional design, providing an international reference for the definition and practice of urban pocket parks.

The classification standards for urban green spaces in China have undergone several revisions in a gradual process of standardization. The 2002 version [24] did not explicitly define the term “pocket park," but reflected related characteristics through two categories of green spaces: “residential area micro-green spaces” and “street-side green spaces”—the former were park green spaces incorporated into residential areas and serving a radius of 300–500 m; the latter referred to independent green spaces outside a road buffer area, emphasizing location attributes and recreational functions. The 2017 version [25] formally added the category of “pocket parks," defined as “independent land-use sites, flexible in size, convenient for daily use by residents, and categorized as medium or small-sized park green spaces". This new standard reclassified residential pocket parks as green spaces, separating them by form into block-shaped parks (no minimum area) and strip-shaped parks (>12 m width), and emphasizing functional-spatial adaptability. Drawing on the Standard for Classification of Urban Green Space (CJJ/T 85-2017) [25] and the Technical Guidelines for Pocket Park Construction (Trial) [26], this paper defines pocket parks as publicly accessible micro-green spaces, typically ranging from 400 to 10,000 m², with recreational functions, distributed within 500 m service radii of communities and streets. While the “pocket park” concept remains fluid and context-dependent, with exact definitions often varying by location, size is widely regarded as the most critical determinant for classification. The definition used here aims to resolve ambiguities in traditional green space classification by emphasizing physical and functional criteria such as block or strip configurations and codified management attributes.

2.2. Study on the Usage Efficiency of Pocket Parks

The usage efficiency of pocket parks is defined as the core manifestation of service effectiveness within unit time and space, specifically measured through dimensions of usage intensity and flexibility. Usage intensity refers to the quantified indicator of user visit frequency per unit area, reflecting spatial carrying capacity. Flexibility indicates the diversity of activities supported by the park, representing the equilibrium and compatibility of resource utilization [7,27].

The usage intensity of pocket parks shows significant spatio-temporal differentiation. Zhou et al. revealed the temporal distribution characteristics of park activities: for example, residents mainly use pocket parks in the evening [9]. When Balai Kerishnan et al. conducted a questionnaire survey of pocket park users in Kuala Lumpur (n = 363), the proportion of short-term stays (≤30 min) was 53.4%, indicating that pocket parks mainly serve the function of short-term rest [8]. Peschardt et al. studied pocket parks in Copenhagen through on-site questionnaires, quantifying usage intensity by recording the actual number of users per unit area (m²) of each pocket park. They found that there was a negative correlation between area and usage density, with small-area parks coming under greater pressure [10]. Cohen et al. used the SOPARC system to evaluate the usage of new pocket parks in low-income communities in Los Angeles. The results suggested that promoting people’s usage of and physical activity in pocket parks cost less than for nearby playgrounds, and the usage frequency of the parks was higher [11].

Usage flexibility focuses on identifying and analyzing the diverse types of activities carried out by park users. Activity types include static rest and restorative activities, dynamic social and group activities, exploratory interactions with nature, daily passage, and short-term stays [9,12,13,14]. For example, according to an empirical study by Ma et al. featuring 72 pocket parks in high-density urban areas in southwest China, the dominant activity types of users were group activities (square dancing, equipment exercise, ball games) and dynamic social and group activities (snack consumption, socializing over chess and cards), supplemented by static rest and exploratory interactions with nature. These present typical characteristics of high facility dependence and low natural relevance [14]. Zhou et al. studied the open spaces in pocket parks in Yancheng and found that the average area of open spaces accounts for a large proportion of the total park area, supporting group activities such as square dancing. They pointed out that open spaces allow residents to organize diverse activities spontaneously, such as impromptu social interactions and temporary fitness and other restorative activities [9].

Building upon previous definitions of pocket park usage efficiency, this paper selected visitor usage frequency and diversity of activity types [14] as indicators of usage efficiency. Usage frequency refers to the total number of visits to the pocket park within a unit of time. Drawing on previous research, pocket park usage frequency was analyzed separately on weekdays and weekends, revealing markedly different usage patterns between these periods [9,10,11,14,28]. Diversity of activity types refers to the number of different behavioral activities performed by users in the pocket park within a unit of time. Studying the diversity of activity types not only describes “what people do in the park” but also provides a reference for analyzing “why certain activities are accepted or rejected". Previous studies have often focused on differences in activity types among different age groups [9,11,14,28]. Such analyses help managers anticipate conflict risks, optimize spatial design, and ultimately achieve a balance between the functions of urban pocket parks as urban green spaces and public leisure spaces [17]. Furthermore, by identifying behavior patterns and influencing factors of different groups, scientific design and management strategies can be developed to reduce health inequalities and enhance participation in physical activities [18].

2.3. Factors Affecting Pocket Park Usage

Existing studies on the factors affecting pocket park effectiveness mainly focus on single-dimensional approaches, highlighting either external factors or internal built-environment elements. External factors encompass macro-level contextual elements that influence a park’s accessibility and appeal to potential visitors [29]. External element-driven paradigms focus on macro-level location features, including key elements such as traffic accessibility, population density, and surrounding land-use mix. For instance, as early as 1959, the gravity model highlighted the nonlinear relationship between accessibility and land-use intensity [29]. A case study in New York in 2021 further confirmed the correlation between residential density and park visit frequency, indicating an exponential increase within a 500 m radius [30]. Additionally, an analysis based on Chicago data in 2020 proved the significant spatial synergy between commercial facilities’ distribution and green space vitality [31]. These studies quantify the impact mechanisms of macro-factors on space usage from different perspectives.

Internal factors constitute the on-site, micro-level attributes of a park’s built environment and management, which collectively have a direct impact on user experience [29]. Internal factor analysis paradigms focus on landscape and facility characteristics of the built environment to analyze their impact on space usage. Regarding landscape element configuration, key dimensions include green view index [3], water feature design [32], and terrain complexity. For example, intimate water design has been proven to improve summer space usage rates effectively [32]. Empirical studies in Kuala Lumpur have shown that micro-topographic changes can enhance spatial exploration [28]. In facility functionality systems, the layout principle of rest facility density should follow the balance of social distance and visual interaction [33]. Lighting comfort involves the nonlinear correlation between light intensity and safety perception thresholds [34]. Maintenance management levels directly affect user loyalty, as evidenced by research in Seoul showing a significant positive correlation between pavement integrity rates and repeat visitation rates [13]. These factors together constitute the internal drivers of spatial vitality.

Existing studies on pocket parks have often analyzed external and internal factors in isolation, focusing narrowly on individual metrics—such as accessibility and population density for external features [29,30,31], or landscape, facilities, and maintenance separately for internal attributes [3,32]. This fragmented approach lacks a multi-tiered indicator system and fails to capture interactions within and across factor types, limiting a holistic understanding of park performance.

To address the research gap presented by the focusing of existing research on either external or internal factors, this study develops a comprehensive framework for evaluating pocket park usage efficiency (see

Table 1. Model of Factors Influencing the Efficiency of Park Use.

Analysis Perspective	Category of Influencing Factors		Number	Influencing Factor Indicators (Measurement)	Refs.
External factors	Macro-factors	Accessibility	A1	500 m service area (ha)	[13,29,34,35,36]
		External environmental characteristics	E1	500 m POI density of residential areas within the service area (individual/ha)	[31]
			E2	500 m Average population density in the service area (person/ha)	[37]
			E3	500 m Diversity of land use in service areas (individual/ha)	[30]
Internal factors	Middle level factors	Layout form	F1	Proportion of recreational activity space (%)	[38]
			F2	Shape index	[39]
	Microscopic factors	Landscape elements	L1	Green space ratio (%)	[12]
			L2	Green view index (%)	[32]
			L3	Plant species richness (type/1000 m2)	[3]
			L4	Water feature ratio (%)	[32]
			L5	Sculpture piece density (individual/1000 m2)	/
		Service facilities	S1	Leisure building area ratio (%)	[15]
			S2	Non-motorized parking lot available? (yes/no)	[15]
			S3	Fitness facility density (individual/1000 m2)	[15]
			S4	Seat facility density (m/1000 m2)	[15]
			S5	Density of signage facilities (individual/1000 m2)	[15]
			S6	Garbage bin density (individual/1000 m2)	[15]
			S7	Is there a public restroom available? (Yes/No)	[15]
		Maintenance and upkeep	M1	Plant maintenance quality (score)	[40]
			M2	Hygiene and cleanliness quality (score)	[28]
			M3	Pavement maintenance quality (points)	[41]
			M4	Facility maintenance quality (points)	[33]

) that integrates 22 influencing factors classified into two primary dimensions (external and internal) and three hierarchical levels (macro, meso, and micro). External factors encompass accessibility and ambient environment characteristics—such as residential point of interest (POI) density, average population density, and land-use diversity within service areas. Internal factors include layout form, landscape elements, service facilities, and maintenance quality, with further specification of micro-level indicators like green view index, facility completeness, and pavement integrity. By explicitly incorporating multidimensional external features and a structured internal classification while emphasizing synergistic inter-factor relationships, this framework provides a more systematic and practical basis for assessing overall usage efficiency of pocket parks.

3. Materials and Methods

3.1. Research Area and Sample Selection

This study focuses on Nanjing’s main urban area (Figure 1a,b), covering districts including Xuanwu, Gulou, Qinhuai, Jianye, Qixia, and Yuhuatai. Based on the “Green Space System Master Plan” (2021–2025) [42] draft catalog, and considering representativeness, spatial distribution balance, typological diversity, and scale variation, we selected 40 pocket park samples (0.8–1.5 ha) from the study area (Figure 1c). The distribution is as follows: Xuanwu (n = 10), Gulou (n = 9), Qinhuai (n = 10), Jianye (n = 6), Qixia (n = 3), and Yuhuatai (n = 2). Sample selection followed four criteria:

• Representativeness: Pocket parks meeting urban greening standards and local planning requirements;
• Spatial balance: Ensuring proportional distribution of green spaces across urban districts, with priority allocation to urban renewal areas;
• Typological diversity: Including both linear (road/river/wall-adjacent, width > 12 m) and compact (street plazas, pocket parks) forms;
• Scale variation: Areas ranging from hundreds of square meters to nearly 1 ha, encompassing diverse internal and external environmental characteristics.

3.2. Data Collection

This study employed behavioral observation methods to conduct field investigations of 40 pocket parks. The surveys were carried out under clear or cloudy weather conditions in autumn and winter, from November 2024 to January 2025, covering both weekdays and weekends. Daily observations were focused on three peak usage periods: morning (9:00–10:00), mid-day (13:30–14:30), and evening (16:30–17:30), with each period lasting one hour to avoid low-usage intervals such as rest and meal-times. Nighttime surveys were excluded for two main reasons. First, the factors affecting nighttime park use—such as lighting, safety, and specific night activities [43]—differ significantly from those at play during the day and would require a separate evaluation framework. This falls beyond the scope of the study. Second, preliminary nighttime data showed low and uneven visitor distribution, resulting in a sample size too small for reliable statistical analysis. Including such data would compromise the validity of the daytime model. Therefore, the findings are strictly applicable to daytime conditions.

Park users were categorized according to China’s standard demographic age groups: infants/children (≤7 years), adolescents (8–17 years), young adults (18–40 years), middle-aged (41–64 years), and elderly (≥65 years). From observation points located in open areas within the parks (to avoid interfering with users), objective recordings of behavioral activities were made using a combination of on-site notes, photography, and videography. Behavioral maps and annotated diagrams were created to collect information on usage time, age distribution, visitor count, activity types, and spatial distribution.

To quantify activity type diversity, we first classified observed behaviors based on Jan Gehl’s categorization of public space activities as necessary, optional, and social [4]. Park users exhibit diverse behaviors, with preferences varying across age groups. Therefore, based on behavioral observations and existing literature [1,10,15,44], activities were grouped by purpose into four categories: Leisure & Recreation (e.g., socializing, meditating, viewing scenery), Exercise & Fitness (e.g., using gym equipment, jogging, group exercises, tai chi), Cultural Entertainment (e.g., group dancing, photography, singing, board games, using children’s play facilities), and Others (e.g., commuting through the park with minimal dwell time). Each category thus encompasses various specific activities. Usage frequency was quantified as the number of users per hour per 1000 m², while activity type diversity was quantified as the number of distinct activity types per hour per 1000 m². The comprehensive usage efficiency metric was calculated using the entropy method, which objectively weights these two indicators (usage frequency and activity type diversity) for an integrated evaluation.

3.3. Model Construction and Data Analysis

To identify the key factors influencing park usage efficiency, we conducted a quantitative analysis of the metrics, as shown in Table 1. We measured these factors through a combination of on-site surveys, behavioral observations, photography, ArcGIS, ENVI, and semantic segmentation.

Among the external influencing factors, the 500 m service area was generated using ArcGIS network analysis. Within this area, residential POI density was calculated using AMap (a leading Chinese digital map service provided by Alibaba) real-estate residential point data. Average population density was derived from the 2020 WorldPop high-resolution population dataset. Land-use diversity was measured by categorizing AMap POI data into functional land types [45] and computing a mixed-use index (1) [46], defined as

$$L=-{\displaystyle \frac{{\displaystyle \sum _{i=1}^N(p_i\log p_i)}}{\log N}}$$

(1)

where $p_i$ is the proportion of the contribution of function $i$ to the total contribution of the plot unit.

In relation to layout form and landscape element, the shape index was calculated according to landscape ecology principles through the formula

$$S={\displaystyle \frac{\mathrm{P}}{2\sqrt{\pi A}}}$$

(2)

where $S$ represents the shape index, $\mathrm{P}$ denotes the perimeter of the green patch, and $A$ signifies the patch area. Green space ratio was calculated based on remote sensing data in ENVI 5.3. Green view index was obtained by taking panoramic photos at high-activity visitor zones and analyzed through semantic segmentation via ADE 20K datasets. Plant species richness was quantified as the ratio of identified plant species to pocket park area through on-site inventories. Water features were analyzed through the ratio of water area and total park area.

The densities of different service facilities (sculpture piece, seat facility, garbage bin, signage facility, etc.) were measured using handheld laser rangefinders and field mapping, and later calculated as the count per unit area. Metrics of maintenance and upkeep dimension were quantified through satisfaction surveys employing a 5-point Likert scale (1–5), where mean scores captured perceived performance across four dimensions: plant maintenance (reflecting pruning/fertilization efficacy on plant vigor), hygiene and cleanliness quality (pertaining to litter and stain removal), pavement maintenance (addressing repair of damaged/unusable surfaces), and facility maintenance (covering restoration of dysfunctional amenities).

These indicators served as independent variables in subsequent Pearson correlation and multiple linear regression analyses, where they were examined alongside dependent variables representing park usage efficiency. This permitted the identification of key factors influencing the usage efficiency of pocket parks. The corresponding statistical analysis results are presented in detail in Section 4.3.

4. Results

4.1. Characteristics of Usage Efficiency on Weekdays and Weekends

Table 2. Daily Usage Frequency of Pocket Parks in Main Urban Area of Nanjing.

Sample Number	Square Measure (m2)	Weekday Usage Frequency (Number of Users/Day/1000 m2)	Weekend Usage Frequency (Number of Users/Day/1000 m2)	Sample Number	Square Measure (m2)	Weekday Usage Frequency (Number of Users/Day/1000 m2)	Weekend Usage Frequency (Number of Users/Day/1000 m2)
1	3293	122.373	141.504	21	2110	118.957	130.806
2	1936	66.632	78.512	22	3301	63.617	66.646
3	3930	30.789	37.913	23	4360	42.202	50.459
4	2794	31.851	44.734	24	5352	32.698	37.930
5	884	101.851	117.695	25	15,784	24.835	27.433
6	3680	107.880	138.043	26	11,781	27.926	32.934
7	8545	17.554	18.841	27	2298	53.035	58.039
8	12,019	10.234	10.816	28	7773	14.923	19.555
9	12,921	101.926	110.903	29	3380	39.645	48.225
10	4541	10.570	17.617	30	2496	40.064	48.878
11	2259	169.967	180.147	31	3680	128.804	128.804
12	4529	20.977	26.056	32	876	27.397	67.352
13	7447	41.764	48.613	33	6240	38.141	40.865
14	9873	11.141	9.420	34	3741	9.088	12.029
15	2783	26.231	18.326	35	2399	64.193	81.700
16	5174	41.556	46.968	36	4357	58.073	61.975
17	2644	23.449	39.713	37	6440	16.149	19.410
18	8430	19.573	11.032	38	3987	10.533	12.038
19	2241	53.540	63.802	39	3767	30.266	32.390
20	7150	17.063	23.776	40	11,834	13.774	16.732

demonstrates significant variation in daily usage frequency across park samples, ranging from 9.088 to 169.967 per 1000 m² on weekdays and from 9.42 to 180.147 per 1000 m² on weekends. Both distributions fall within the 9–200 range, with greater dispersion observed during weekends. The vast majority of samples recorded usage frequencies below 80. On weekdays, 33 samples (82.5%) fell into the 0–80 range, six (15%) were between 80–140, and only one (2.5%) exceeded 140. The following parks demonstrated higher usage rates per unit time and area with frequencies above 80: sample 1 (Mei’an Garden on Beijing East Road), sample 5 (Houzaimen Street Side Garden), sample 6 (Municipal Government Street Side Garden), sample 9 (Zhenzhu River Linear Garden), sample 11 (Mochou Technology Square Garden), sample 21 (Hongxing Square Garden), and sample 31 (South Lake Pocket Garden). Among these, Mochou Technology Square Garden had the highest usage frequency on weekdays. On non-workdays, 32 samples (80%) were in the 0–80 range, six (15%) in the 80–140 range, and two (5%) above 140. In addition to the parks listed above, Jixian Street Pocket Park (sample 35) also exhibited high usage intensity. Mochou Technology Square Garden again showed the highest frequency on non-workdays. Overall usage frequency patterns remain comparable between weekdays and weekends, though weekends exhibit moderately higher usage. The original fieldwork observation data can be found in Supplementary Table S2.

As shown in

Table 3. Diversity of Daily Activity Types.

Sample Number	Square Measure (m2)	Diversity of Types of Activity on Weekdays (Types/Day/1000 m2)	Diversity of Types of Activity on Weekends (Types/Day/1000 m2)	Sample Number	Square Measure (m2)	Diversity of Types of Activity on Weekdays (Types/Day/1000 m2)	Diversity of Types of Activity on Weekends (Types/Day/1000 m2)
1	3293	13.665	8.502	21	2110	27.488	28.910
2	1936	14.979	19.628	22	3301	9.694	11.512
3	3930	11.196	12.214	23	4360	8.257	9.174
4	2794	11.810	15.388	24	5352	8.221	8.782
5	884	36.214	39.609	25	15,784	2.788	3.104
6	3680	17.663	11.957	26	11,781	4.753	5.432
7	8545	2.926	2.809	27	2298	11.007	12.008
8	12,019	2.829	3.079	28	7773	3.988	4.631
9	12,921	5.263	6.888	29	3380	16.568	18.935
10	4541	5.726	9.249	30	2496	7.212	11.218
11	2259	28.770	29.656	31	3680	10.598	13.315
12	4529	8.832	9.274	32	876	15.982	27.397
13	7447	5.372	5.506	33	6240	6.891	7.051
14	9873	1.317	1.418	34	3741	5.079	9.623
15	2783	3.953	8.983	35	2399	17.090	18.758
16	5174	9.084	9.858	36	4357	6.886	7.804
17	2644	3.782	14.750	37	6440	5.435	6.366
18	8430	1.542	2.610	38	3987	5.517	6.270
19	2241	17.847	18.739	39	3767	10.089	10.620
20	7150	4.336	6.014	40	11,834	3.803	4.225

, daily activity type diversity varies considerably among the park samples. Weekday diversity ranges from 1.317 to 36.213, while weekend diversity ranges from 1.418 to 39.609, with both distributions falling between 1 and 40 and showing greater variation on weekends. The majority of samples (80%, n = 32) exhibit weekday diversity scores between 0 and 15, while 17.5% (n = 7) fall within the range 15–30, and only 2.5% (n = 1) exceed 30. Similarly for weekends, 77.5% of samples (n = 31) show diversity scores within the range 1–15, 20% (n = 8) fall within the range 15–30, and 2.5% (n = 1) exceed 30. Samples exceeding 15 demonstrate significantly higher activity type diversity per unit time/area. Collectively, activity diversity patterns remain comparable between weekdays and weekends, though weekends show moderately higher diversity. The original fieldwork observation data can be found in Supplementary Table S1.

These findings demonstrate temporal influences on park usage. Regarding visitation frequency, weekends show higher usage rates with more diverse user demographics compared to weekdays. Activity type diversity also increases on weekends, featuring more varied behaviors and greater participation in group activities.

Drawing on the calculations presented in Table 2 and Table 3 and visualized in Figure 2, we employed a paired-samples t-test to analyze differences in usage behavior between weekdays and weekends among 40 parks. Descriptive statistics showed that pocket park usage frequency was higher on weekends (M = 54.47, SD = 43.14) than on weekdays (M = 47.03, SD = 39.24). Activity diversity was also greater on weekends (M = 11.53, SD = 8.30) compared to weekdays (M = 9.86, SD = 7.70). Correlation analysis revealed a strong positive relationship between weekday and weekend usage frequency (r = 0.982, p < 0.001) and activity diversity (r = 0.931, p < 0.001). The t-test results further indicated that usage frequency was significantly higher on weekends than on weekdays (mean difference = −7.44, t (39) = −5.36, p < 0.001), as was activity diversity (mean difference = −1.67, t (39) = −3.50, p = 0.001). These results suggest a significant weekend effect in relation to both user behavior and activity diversity.

4.2. Comprehensive Evaluation of Park Usage Efficiency

Entropy-based weighting analysis (

Table 4. Weight of Efficiency Indicators.

Time	Usage Frequency	Use Frequency Index Weights	Diversity of Activity Types	Weight of Diversity Indicators for Activity Types
Weekday	w1	0.5741	w2	0.4259
Weekend	w1	0.5885	w2	0.4115

) reveals that usage frequency carries greater weight than activity type diversity for both weekdays and weekends, indicating its stronger influence on park efficiency. Higher visitation frequency signifies greater user numbers, while elevated activity diversity expands behavioral choices. This combination prolongs visitor engagement and attracts broader participation. These relationships collectively demonstrate that parks accommodating more users inherently facilitate greater variety in activity generation.

Comprehensive usage efficiency scores range from 0 to 0.1 across park samples (Figure 3). Based on central tendency analysis, evaluation outcomes were categorized into four tiers: excellent (A: 0.07–0.1), good (B: 0.04–0.07), medium (C: 0.01–0.04), and poor (D: 0–0.01). Weekdays show three A-level samples (7.5%), four B-level (10%), 19 C-level (47.5%), and 14 D-level (35%). Weekends show two A-level samples (5%), seven B-level (17.5%), 20 C-level (50%), and 11 D-level (27.5%). These distributions indicate generally moderate overall efficiency, with weekends demonstrating slightly better performance, as evidenced by more C-tier-and-above parks (72.5% vs. 65%). However, the majority of samples remain concentrated in the C and D tiers (82.5% weekdays, 77.5% weekends), suggesting marginal performance across most locations.

4.3. Factors Affecting Usage Efficiency: Correlation and Regression Analysis

A collinearity test was conducted prior to regression analysis to diagnose the severity of multicollinearity among variables in the multiple linear regression process. Higher VIF values indicate stronger collinearity between variables [47], meaning they exhibit high linear correlation, which may compromise the accuracy of regression results. Generally, a VIF below 10 suggests no significant multicollinearity; a VIF between 10 and 100 indicates moderate multicollinearity; and a VIF of 100 or higher signals severe multicollinearity. Our test results showed no significant multicollinearity, confirming the suitability of multiple linear regression for modeling.

Using SPSS Statistics 27.0, we employed Pearson correlation coefficients to assess relationships between different indicators and park usage efficiency (

Table 5. Correlation Analysis between Usage Efficiency and Influencing Factor Indicators.

Weekday
factor	A1	E1	E2	E3	F1	F2	L1	L2	L3	L4	L5
correlation coefficient	0.371 *	0.160	0.085	0.132	0.609 **	−0.087	−0.337 *	0.039	0.699 **	−0.031	0.000
factor	S1	S2	S3	S4	S5	S6	S7	M1	M2	M3	M4
correlation coefficient	0.606 **	−0.205	0.295	0.645 **	0.000	0.382 *	−0.087	0.630 **	0.458 **	0.269	0.288
Weekend
factor	A1	E1	E2	E3	F1	F2	L1	L2	L3	L4	L5
correlation coefficient	0.333 *	0.153	0.053	0.185	0.573 **	−0.103	−0.308	0.052	0.761 **	−0.073	0.139
factor	S1	S2	S3	S4	S5	S6	S7	M1	M2	M3	M4
correlation coefficient	0.647 **	−0.244	0.258	0.654 **	0.071	0.444 **	−0.152	0.667 **	0.482 **	−0.061	−0.296

Note: * Significant correlation at the 0.05 level (bilateral); ** Significant correlation at the 0.01 level (bilateral).

). For weekdays, significant correlations emerged at p < 0.01 between usage efficiency and the following variables: F1 (Proportion of recreational activity space), L3 (Plant species richness), S4 (Seat facility density), and M2 (Hygiene and cleanliness quality). Significant correlations at p < 0.05 existed with A1 (500 m service area), L1 (Green space ratio), S6 (Garbage bin density), and M1 (Plant maintenance quality). On weekends, significant correlations at p < 0.01 were observed with F1, L3, S1, S4, S6 (Garbage bin density), M1 (Plant maintenance quality), and M2 (Hygiene and cleanliness quality), while A1 (500 m service area) showed significance at p < 0.05. The highest correlation coefficient (L3 = 0.761 on weekends) remains below the 0.9 multicollinearity threshold, ensuring the validity of subsequent multiple regression analyses.

The regression results indicate adjusted R² values of 0.695 and 0.762 for the two models, respectively, with both demonstrating reliability per F-tests (

Table 6. Results of Multiple Linear Regression.

Weekday
Independent Variable	Non Standardized Coefficient		Standard Coefficient	t	Sig	Collinearity Test
	Beta	Standard Error				Tolerance	VIF
Constant	−0.052	0.017		−3.149	0.004
A1 (500 m service area)	0.000	0.000	0.135	1.341	0.190	0.772	1.295
F1 (Proportion of recreational activity space)	0.031	0.018	0.222	1.758	0.089	0.491	2.038
L1 (Green space ratio)	0.013	0.015	0.100	0.864	0.394	0.587	1.704
L3 (Plant species richness)	0.001	0.000	0.329	2.872	0.007 ***	0.594	1.683
S4 (Seat facility density)	0.000	0.000	0.246	2.097	0.044 **	0.566	1.766
S6 (Garbage bin density)	0.000	0.003	−0.034	−0.330	0.743	0.737	1.357
M1 (Plant maintenance quality)	0.005	0.003	0.206	1.804	0.081 *	0.600	1.668
M2 (Hygiene and cleanliness quality)	0.005	0.002	0.219	2.210	0.035 **	0.795	1.258
R	0.871
R2	0.758
Adjust R2	0.695
DW value	1.949
F	12.127
F significance	0.000
Weekend
Independent variable	Non standardized coefficient		Standard coefficient	t	Sig	Collinearity test
	Beta	Standard Error				Tolerance	VIF
Constant	−0.033	0.009		−3.743	0.001
A1 (500 m service area (ha))	0.000	0.000	0.082	0.927	0.361	0.779	1.284
F1 (Proportion of recreational activity space)	0.012	0.012	0.097	0.978	0.335	0.627	1.596
L3 (Plant species richness)	0.001	0.000	0.427	4.208	0.000 ***	0.594	1.685
S4 (Seat facility density)	0.000	0.000	0.181	1.805	0.080 *	0.605	1.652
S6 (Garbage bin density)	0.001	0.002	0.035	0.389	0.700	0.740	1.352
M1 (Plant maintenance quality)	0.005	0.002	0.235	2.331	0.026 **	0.602	1.662
M2 (Hygiene and cleanliness quality)	0.005	0.002	0.219	2.498	0.018 **	0.793	1.261
R	0.897
R2	0.897
Adjust R2	0.762
DW value	1.934
F	18.798
F significance	0.000

Note: * indicates significant differences; ** indicates highly significant differences; *** indicates very significant differences. Sig < 0.01 indicates that the difference is very significant; 0.01 < Sig < 0.05 shows significant difference; 0.05 < Sig < 0.1 shows marginally significant difference; Sig > 0.1 indicates no significant difference.

). As this study aims to establish a non-predictive model for exploring significant variable impacts on utility with limited sample size, significance criteria were appropriately relaxed and set at the 0.1 level. Subsequent significance analysis therefore references the same criteria.

For landscape elements, L3 (Plant species richness) demonstrates a highly significant positive correlation with park usage efficiency (p < 0.01, β = 0.329), indicating that enhanced floral diversity substantially improves functional performance. This relationship confirms that visitors prioritize environmental quality, as richer plantings significantly encourage visitation.

For service facilities, S4 (Seat facility density) demonstrates statistically significant positive effects at p < 0.05 on weekdays and p < 0.10 on weekends (β = 0.246, ranking second among the significant factors). This confirms that increasing seating provision enhances spatial efficiency.

For maintenance management, M2 (Hygiene and cleanliness quality) demonstrates significant positive correlation (p < 0.05, β = 0.219), confirming that improved hygiene standards effectively enhance park efficiency. The regression coefficient ranks highly among the significant factors. Clean parks consistently attract visitors by providing fundamental recreation conditions. Meanwhile, M1 (Plant maintenance quality) approaches significance (weekdays: p < 0.10; weekends: p < 0.05, β = 0.206), suggesting enhanced vegetation conditions moderately improve efficiency. Poorly maintained vegetation diminishes environmental quality and visitor experience, reducing both enjoyment and visitation willingness.

Regarding spatial configuration, F1 (Proportion of recreational activity space) demonstrates marginal significance (p < 0.10, β = 0.222), suggesting that increased recreational allocation may enhance site efficiency. This aligns with parks’ primary recreational function, as activity spaces fundamentally enable diverse outdoor usage patterns.

Indicators not reaching statistical significance in this study may still influence the dependent variable. For accessibility metrics, the 500 m service area shows borderline significance (p = 0.19, β = 0.135), suggesting park accessibility improvements represent potential optimization opportunities, despite falling short of conventional thresholds.

5. Discussion

5.1. Usage Patterns in Different Pocket Parks

While the overall trend of urban park usage is similar across the week, this study reveals pronounced differences on weekends, which exhibit markedly greater visit frequency and a broader diversity of activities compared to weekdays. This trend is in line with the “weekend leisure” pattern as observed by Gao et al. [48]. Chinese park users display clear generational differences: Core users (seniors and children) visit parks consistently throughout the week, while occasional users (working-age adults) concentrate visits on weekends [49]. Free from work constraints, core users maintain relatively fixed activity schedules with minimal weekday–weekend variation [50]. Furthermore, visitor activity diversity within parks may relate to demographic factors like age and gender. These contrasting usage patterns reflect how different social groups allocate time in transitioning societies and have implications for planning and design in diverse socio-cultural contexts [51].

From the perspective of spatial justice [49], the efficiency gaps reflect deeper inequities: traditional planning and design metrics like 500 m service radius only address physical accessibility, ignoring generational variations in time availability. This gap becomes evident in conflicting time demands—such as seniors’ need for morning microclimate comfort versus working-age adults’ evening activity requirements and weekday–weekend usage overlaps—revealing fundamental spatiotemporal mismatches. To resolve this, we propose the adoption of “temporal design,” enabling single spaces to serve multi-timeframe functions [52], for example, in the mornings for age-friendly zones, in the daytime for commuter services, and in the evenings for youth activation. Thus, limited physical areas can be transformed into adaptable community assets through intelligent time-based resource allocation.

Furthermore, using entropy methods, we aggregated and compared the usage efficiency of different pocket parks. Most pocket parks scored at medium or low (Bands C and D) levels on weekdays. While weekend performance showed a slight improvement, it remained generally at a medium level (Band C and above). This underperformance may have stemmed from two key obstacles in Nanjing’s pocket parks: limited physical space and improper management. Inadequate land allocation restricts functional diversity and user capacity; and insufficient maintenance budgets and uncoordinated operational protocols lead to declining facility quality, chaotic public order, and ultimately a diminished user experience. These findings highlight significant room for improvement in the design and operation of the city’s pocket parks. By identifying the causes of these efficiency gaps, we can better target upgrades where they matter most, as examined in the next subsection.

5.2. Factors Influencing Pocket Park Usage

The results classified different independent variables into three levels: statistically significant factors in the regression, factors not statistically significant but correlated, and uncorrelated factors. Statistically significant factors include L3 (Plant species richness), S4 (Seat facility density), M1 (Plant maintenance quality), and M2 (Hygiene and cleanliness quality). Regarding diversity and management of plants, according to the biophilia hypothesis, humans have an innate tendency—seen as the product of biological evolution—to seek a connection with nature [53,54]. The diversity and maintenance level of parks may lead to enhanced perception of aesthetic and ecological values, in turn increasing frequency of visits as well as enhancing diversity of activities.

Elevated amenity (e.g., seat facility) density substantially enhances recreational capacity and rest opportunities [55], directly extending public dwell time and diversifying activity typologies [4]. Further, sanitation standards serve as visible indicators of managerial investment [56], with empirical studies confirming their status as primary predictors of user satisfaction [57]. This may also result in a significant influence on park usage, both on weekdays and weekends. The proportion of recreational activity space demonstrates marginal significance; while optimizing this proportion provides key leverage for improving utilization rates, maximizing coverage proves counterproductive, as strategic land allocation balancing recreational and complementary functions yields optimal outcomes.

Several factors exhibited a correlational tendency but did not reach statistical significance. These included A1 (500 m service area) r = 0.371 on weekdays, r = 0.333 on weekends, L1 (Green space ratio) (r = −0.337 on weekdays), and S6 (Garbage bin density) (r = 0.382 on weekdays, and 0.444 on weekends). Specifically, the lack of significant effect for service area size may be due to marginalization in high-density urban settings, where facility catchments widely overlap and accessibility is saturated, diminishing the role of service area size as a differentiating variable. This aligns with Guzman et al. [58], who found that beyond a certain density, proximity exhibits diminishing returns. Therefore, our results do not contradict but rather refine previous understanding, suggesting that the influence of such variables depends on contextual conditions, qualitative attributes, and the presence of stronger predictors.

Similarly, neither the green space coverage ratio nor the green view index demonstrated a significant positive relationship. This lack of association for the green view index contrasts with prior findings showing its beneficial role in restoring people’s perception of their environment [59,60]. One reason may be that if a pocket park is “green” by the data (high coverage/view index), but has poor design, overgrown plants, or bad upkeep, or feels unsafe, this visual greenness will fail to attract users. As Zhang et al. [61] argue, the perceived quality, maintenance, and safety of green areas are stronger predictors of usage than the simple ratio metric used in urban planning. The results regarding garbage bin density align with the work of Smith et al. [62], who found that mere proximity to bins was equally as important as clear signage and public education campaigns in promoting proper waste disposal. This suggests that the quality and accessibility of infrastructure may be a more salient factor than its sheer quantity.

Another surprising finding was that external environmental features (land-use density, residential area, and population density within 500 m) were not correlated with pocket park usage efficiency. This finding conflicts with transportation-planning models developed by Cervero and Kockelman (1997) [63] in their 3D (Density, Diversity and Design) framework, but substantiates emerging critiques about multimodal indicator miscalibration identified by Boisjoly and El-Geneidy (2017) [64].

Methodological considerations may explain several non-significant associations in this research. The scale mismatch hypothesis proposed by Wu (2004) [65] in urban ecology studies appears applicable to our minor factor analyses, where micro-scale elements (e.g., decorative features such as pavements, sculpture, and service facilities) demonstrate attenuated effects in large-level assessments. This attenuation directly stems from the discordance between localized interventions and macro-level assessment frameworks. Consequently, our findings substantiate the need for multi-scalar approaches as advocated by Pickett et al. (2008) [66] in urban ecosystem research.

5.3. Theoretical and Practical Implications

This study builds upon traditional research that often focuses narrowly on isolated aspects of park usage efficiency, attaining no comprehensive understanding of indicator selection or the construction of influencing factor models. Building upon empirical studies of pocket parks in Nanjing, it constructs an influencing factor model at macro-, meso-, and micro-scales covering both internal and external aspects. It quantifies built-environment characteristics of parks alongside usage efficiency metrics, and employs correlation and regression analyses to investigate how built-environment determinants affect park usage on weekdays versus weekends. By further identifying the relative impact of these factors, the research provides guidance for the scientific optimization and enhancement of urban park design. In this way, this research transforms abstract spatial equity principles into an evidence-based design paradigm, providing a scalable, context-responsive model for optimizing urban green infrastructure.

Building upon the analysis of influencing factors and stratification of park usage efficiency, this study proposes optimization strategies for pocket park renewal within planning and design disciplines. We assessed comprehensive usage efficiency across all parks via entropy methods and categorized them into distinct performance tiers. This stratification establishes urban renewal priorities, indicating that parks in the lowest efficiency tier warrant prioritized intervention.

Five priority strategies target critical factors—specifically, Proportion of recreational activity space (F1), Plant species richness (L3), Seat facility density (S4), Plant maintenance quality (M1), and Hygiene and cleanliness quality (M2)—to enhance park efficiency. For example, local authorities should balance green-to-hardscape ratios by reducing ornamental green spaces and prioritizing recreational areas, thereby optimizing land resources and fostering social interaction. Ecologically, planting schemes should center on native species for resilience, supplemented by seasonal ornamentals in focal areas; concurrently, vertical greening and multi-layered communities can enhance biodiversity cost-effectively. Amenity management requires dynamic seating adjustments based on pedestrian flow using modular designs (e.g., movable benches, stepped seating), strategic waste bin placement aligned with activity density, and implementation of water features, sculptures, and restrooms to avoid overdesign.

Secondary strategies can utilize key indicators such as A1 (500 m service area), L1 (Green space ratio), and S6 (Garbage bin density) for prioritization. These evidence-based recommendations seek synergistically to enhance ecological integrity, social inclusivity, and operational feasibility within urban green systems. It is also important to note that statistically insignificant factors may still hold practical relevance and can be implemented as supplementary measures.

6. Conclusions

Pocket parks, as compact yet vital components of urban green space systems, are particularly crucial during urban renewal where land is scarce. This study pioneers a comprehensive assessment of pocket park efficiency by combining internal design and external environmental factors. Fieldwork conducted across 40 parks in Nanjing’s main area on both weekdays and weekends revealed that (1) visitation frequency and activity diversity are significantly higher on weekends; (2) higher visitation frequency correlates with greater activity diversity; and (3) most parks exhibit moderate overall usage efficiency, indicating a need for systematic enhancement. Statistical analysis identified key influencing factors, including the proportion of recreational activity space, plant species richness, seating facility density, plant maintenance quality, and hygiene and cleanliness quality. Based on these findings, targeted planning and renewal strategies were proposed.

Limitations of this study lie in the following aspects. First, this study focused on daytime park use in autumn and winter. Consequently, the findings may not fully represent nighttime usage patterns or seasonal behaviors in spring and summer, which could limit generalizability. Second, the study primarily employed on-site observation for usage efficiency assessment. While this method provides firsthand data, it may introduce observer bias and hinder the scalability of data collection. Future research could integrate drone orthophotography with crowd density algorithms and Wi-Fi probes to enable unobtrusive observation of visitor behavior. Third, the statistical model excluded socioeconomic and microclimatic variables. This omission may affect the explanatory power of the model, as factors such as income levels or local weather conditions could also influence park usage. Future studies could incorporate structural equation modeling along with these omitted variables to examine potential mediation pathways. Fourth, the study’s geographical focus is limited to Nanjing’s unique urban environment, characterized by high population density and specific cultural habits. While the findings may have applicability to similar high-density Asian cities, they may be of lesser relevance to European or North American cities with different urban structures, cultural norms, and mobility patterns. Future comparative studies across diverse urban settings are needed to establish the boundary conditions of these patterns.