Next Article in Journal
Healthy Campus: A Contribution to the Environment, Sustainability, and Social Responsibility Practices at the IPBeja in Portugal
Previous Article in Journal
Using Virtual Reality to Enhance Learning Performance and Address Educational Resource Disparities in Architectural History Courses
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Leveraging City Cameras for Human Behavior Analysis in Urban Parks: A Smart City Perspective

1
Faculty of Architecture and Town Planning, Technion-IIT, Haifa 3200003, Israel
2
Faculty of Electrical and Computer Engineering, Technion-IIT, Haifa 3200003, Israel
*
Author to whom correspondence should be addressed.
Sustainability 2025, 17(3), 865; https://doi.org/10.3390/su17030865
Submission received: 7 November 2024 / Revised: 7 January 2025 / Accepted: 10 January 2025 / Published: 22 January 2025

Abstract

:
Urban parks enhance urban life by providing essential spaces for recreation, relaxation, and social interaction. However, there is a lack of understanding of how park settings influence usage patterns by socio-demographic characteristics. This study seeks to address this gap by exploring the association between park characteristics and gendered usage patterns across different times of the day. We employed big data analytics and computer vision techniques to analyze human behavior in two urban parks. These parks have comparable environments characterized by shared features, including paths, playgrounds, seating, lawns, greenery, and amenities. One is designed as a linear park, while the other is trapezoid-shaped. The distribution of facilities varies within the parks’ spaces. The key innovation of this approach lies in the use of computer vision for spatial analysis based on user-specific characteristics, particularly gender. City surveillance cameras are leveraged to gather extensive data on park usage. A comparative evaluation of the two urban parks includes a detailed examination of temporal and spatial usage patterns, offering new insights into the dynamics of urban park utilization. Findings reveal specific park features, such as playgrounds and paths, showed varying levels of utilization by different genders, highlighting the importance of tailored urban design. Males favored open lawns with dog facilities, whereas females preferred areas near playgrounds. The application of smart city technologies, such as city cameras, sets the stage for future directions in urban planning and design, highlighting opportunities to integrate advanced analytics into planning practices.

1. Introduction

The concept of smart cities and innovative urban technologies has significantly influenced urban planning and development [1]. These technologies, which include data collection, computing, and communication, are crucial for addressing urban challenges and improving city dwellers’ quality of life [2]. Urban planners are increasingly engaging with smart city processes, with digitalization profoundly impacting their work [3]. Key technologies such as high-performance computing, Internet of Things (IoT) sensors, smart power grids, Geographic Information Systems (GIS), Global Positioning Systems (GPS), Radio Frequency Identification (RFID), Augmented Reality (AR), and Virtual Reality (VR) are crucial in optimizing urban planning and infrastructure management [4].
Various data sources have been employed to study human activities, including mobile phone network data, GPS data, and social media, all of which facilitate the collection of spatial and temporal information [5]. Some studies have leveraged social media platforms and online reviews to assess visitor experiences and usage in public open spaces [6,7,8,9]. Although social media and online review data provide valuable spatial and demographic information, the self-reported nature of demographic details can be misleading, as users may not always provide accurate information. Tracking and sensing data, such as Bluetooth sensors and GPS data, are widely used to study pedestrian spatial behavior [10,11]. While these sources offer rich spatial and temporal data, they often lack demographic information.
Recent advancements in computer vision and deep learning have enabled the use of camera imagery data, such as Google Street View and Tencent Maps, to assess human perceptions in urban spaces [12,13,14]. These images, freely available to the public, capture urban spaces from users’ viewpoints, facilitating the evaluation of user perceptions without the need for on-site visits, which saves time. However, these data sources have significant limitations: They lack temporal attributes, are restricted to a single scene at a specific time, and do not provide demographic information. Unmanned Aerial Vehicles (UAVs) have also been used to collect park usage data [15]. They have several limitations, including the need for skilled personnel, high costs, and susceptibility to weather conditions.
Studies have investigated the application of cameras in assessing human behavior, demonstrating the significant potential of video data from various sources. Tay [16] explores the utility of video data for behavioral analysis, while Yamamoto et al. [17] examine customer behavior using top-view depth cameras. Antonakaki et al. [18] focus on methodologies for detecting abnormal behavior through multi-camera systems, and Rezaei et al. investigate vision-based techniques for classifying human actions [19]. These studies highlight the promise of camera technology in capturing and analyzing human behavior with various applications. Moreover, the review by Ibrahim et al. [20] identifies significant opportunities for using deep learning and computer vision models to enhance the precision and efficiency of human behavioral assessments in cities.
Most studies evaluating human activity in urban parks rely on questionnaires [21,22] and observations [23]. The most common tool for evaluation was employed through either on-site auditing observation [23], a walking tour [24], or remote observation using satellite imagery from Google Earth [25]. Another significant observation tool is behavioral mapping, which is particularly effective for assessing human behavior in urban parks, as it considers both temporal and spatial dimensions alongside user socio-demographic characteristics. When combined with GIS, behavioral mapping becomes a powerful method for observing and recording human activities at specific times within given environments [26,27]. Despite its advantages, behavioral mapping has its limitations. Research using this method often involves relatively small sample sizes. Moreover, behavioral mapping to collect spatial and temporal data necessitates prolonged observation over an extended period, highlighting its time-consuming and demanding nature. Given the limitations of existing data sources, there is an urgent need to explore new research methods.
Despite the capabilities of existing tools, they fall short in examining the interaction between users’ socio-demographic characteristics and urban park features, particularly when incorporating temporal and spatial dimensions grounded in comprehensive data collection. To address this gap, this study aims to explore the association between park characteristics and gendered usage patterns across different times of the day. We analyze human behavior in urban parks through the utilization of big data. Specifically, it focuses on a comparative evaluation of two urban parks, leveraging data obtained from Closed-Circuit Television (CCTV) cameras. The methodology integrates traditional geospatial analysis with advanced computer vision techniques, offering a novel framework for data collection and analysis in urban park usage research. Based on empirical data, the study introduces an analytical approach to urban park use, enabling urban planners and managers to make more informed decisions. This approach contributes to the achievement of SDG 11: Sustainable Cities and Communities [28] by facilitating the optimization of land uses, enhancing urban environmental quality, and fostering social inclusion.

2. Background and Related Work

2.1. Human Activity in Urban Parks

Historically, Public Open Space (POS) has accommodated diverse human activities, serving as sports activity spaces [29], therapeutic environments [30], commercial hubs [31], areas for staying/moving between buildings [32], spaces for artistic expression [33], and spaces for protest [34]. In contemporary contexts, the demand for outdoor activities has surged due to changing lifestyles and rising living standards, heightening the importance of POS [35]. Gehl [32] describes POS as “life between buildings”, encompassing all outdoor areas within the built environment, including public and private domains. POS can be broadly classified into two primary categories. The first category encompasses staying areas for conducting activities, such as parks, gardens, plazas, playgrounds, and sports fields. The second category pertains to areas intended for passage, facilitating movement from one location to another. These areas include streets and pedestrian zones [36].
In this study, we focus on urban parks since they play a vital role in the neighborhood’s urban fabric [37]. Urban parks offer city dwellers abundant outdoor leisure and recreation opportunities [38]. They also serve as accessible recreational spaces during stressful times when other recreation options are limited [8] and provide a range of physical and psychological benefits to city dwellers, which was particularly evident during the COVID-19 pandemic [39,40]. The quality of urban parks holds greater significance in delivering beneficial effects to city residents [41]. Studies have demonstrated that proximity to urban parks is significantly associated with improved mental health outcomes, as evidenced by higher scores on the Mental Health Index (MHI) [42]. Furthermore, the mental health benefits of parks extend to both residential areas and workplace environments [43]. These benefits include stress reduction, improved mental health, and increased social interaction [39,40,44,45]. Urban parks have been found to have environmental, economic, and social benefits, further enhancing their value to city residents [41]. Studies have extensively explored the influence of urban park features on various activities, particularly physical activities [46,47,48] and social activities [49]. Moreover, the physical characteristics of the surrounding environment have been examined for their impact on human behavior [26,32,50,51,52,53]. For instance, Goličnik and Thompson [26] analyzed park design elements, revealing that components such as grassed areas, landmarks, and edges significantly influence the utilization patterns of these spaces. These studies highlight the crucial role of intentional urban park design in promoting diverse activities and enhancing user experiences. This study focuses on neighborhood urban parks. Since they offer convenient and nearby options for leisure and recreation, they play a crucial role in residents’ daily lives.

2.2. Influence of Socio-Demographic Characteristics on Patterns of Park Usage

Research on park usage and its relationship to socio-demographic characteristics has yielded several findings regarding human behavior in urban parks. Park proximity, facilities, and neighborhood characteristics were found as factors influencing park use, with variations observed across gender, age, race, and income groups [54,55]. McCormack et al. [56] and Arısoy [57] emphasized the importance of safety, aesthetics, amenities, and maintenance in promoting park usage. Their research indicates that occupation, income, and age significantly determine how often and why individuals visit parks. This highlights the multifaceted nature of park use, suggesting that improvements in park design and maintenance could enhance user experiences and encourage more frequent visits.
Studies by Ma et al. [58] and Cohen et al. [48] have shown that socio-demographic factors such as gender, age, education level, and income play a crucial role in influencing park usage and preferences. These findings suggest that targeted interventions and policies are necessary to address the specific needs of different demographic groups, thereby promoting equitable access to urban parks. Other studies have examined the relationship between gender and time in park usage [59,60,61,62,63] and between gender and specific park zones [27,61,64,65,66]. For instance, Marušić [27] discovered that urban parks experience varying occupancies depending on the time of day. This study highlighted that park features influence activities based on gender and age; for example, young males predominantly use skateboarding areas. These research findings indicate that the usage of urban parks varies according to demographic characteristics and specific times of the day and week.
Other studies highlight that environmental and social factors significantly influence perceptions of safety in urban parks, which impacts their usage. Women consistently prioritize features that enhance visibility and minimize the risk of harassment, such as adequate lighting, well-maintained pathways, and open designs with clear sightlines, as well as the integration of spaces within neighborhoods and the presence of visible security measures like cameras and patrols [67]. Similarly, another study highlights that women value functional infrastructure and well-maintained greenery and are more affected by the presence of disruptive individuals, such as those consuming alcohol [68]. For men, safety perceptions are shaped more by environmental cues such as the condition of the space, the presence of escape routes, and active community zones, with less emphasis on maintenance and social disruptions [67,68]. Both genders, however, share common priorities, such as lighting, highlighting the factors that influence the use of urban spaces. These studies suggest that while the core aspects of safety perception overlap, the nuanced differences between men and women should guide urban design to cater to diverse user needs effectively.
This study aims to analyze gender-based usage patterns of comparable urban park environments with distinct designs and layouts. One park features a linear and compact layout, while the other has a trapezoidal shape and a dispersed layout. The research focuses on the distribution of activity zones within these parks, examining how these spatial arrangements influence usage patterns. By comparing the similar features (such as paths, playground areas, greenery, etc.) of two parks, we seek to understand how these elements are utilized within an urban park setting and how they cater to different user groups.

3. Methodology

In this study, we employ a comparative evaluation method to identify the park features associated with varying activity levels between genders in two urban parks. These parks are located in different neighborhoods within the same city of Or-Yehuda, Israel, and share similar socioeconomic characteristics among the neighborhoods’ residents. Both parks have common features such as paths, seating areas, playground equipment, greenery, lawns, and amenities. However, one (Remez-Garden) is designed as trapezoid-shaped and offers unique activities, such as a dog facility, while the other (Katznelson-Garden) is designed as a linear park. The evaluation aims to compare user activities in relation to these similar and distinct features by gender. This approach allows for a detailed understanding of how specific features contribute to gendered patterns of park use, offering implications for urban planners and designers in creating inclusive and engaging parks.

3.1. Study Area

Or-Yehuda is a moderately sized city (with approximately 40,000 inhabitants) situated within the metropolitan area of Tel-Aviv, Israel. It has 33 urban parks, whereas 25 parks are smaller than 5000 square-meters.
These parks were selected based on several criteria: size, shape, diversity of facilities and activity areas, and the presence of CCTV cameras. To optimize surveillance coverage, we selected parks that encompass large areas with extensive camera monitoring. The chosen parks are Katznelson-Garden and Remez-Garden. They are located in different neighborhoods, yet share a similar socioeconomic index. The Israeli-Central-Bureau-of-Statistics (CBS) created the socioeconomic index to rank and compare municipalities and neighborhoods based on the socioeconomic level of their populations. This index relies on 14 variables distributed across four primary domains: demography, education, employment and retirement, and standard of living, which includes aspects like financial income, satisfaction levels, and housing conditions. The index ranges from one to ten, where one signifies the lowest socioeconomic status and ten the highest [69]. These parks were chosen for their similar characteristics, including playground areas, lawns, greenery, paths, and their surroundings, including residential buildings and kindergartens. Both serve as neighborhood parks. However, they differ by shape—Katznelson-Garden is a linear park, while Remez-Garden is a trapezoid-shaped park. Katznelson-Garden is surrounded by multiple parks and is well-connected with pathways, whereas Remez-Garden lacks nearby parks in its neighborhood. Although the cameras in Remez-Garden may seem clustered in a nearby location, their angles are carefully positioned to focus on distinct activity zones. In Katznelson-Garden, which is narrower and more elongated, the cameras were installed in separate locations to effectively cover the park’s various activity zones (Figure 1 and Figure 2). This placement strategy ensures that in both parks, the cameras comprehensively capture the primary areas of visitor activity.
Katznelson-Garden is situated in the Degania neighborhood on the southeast side of Or-Yehuda, home to 3505 residents and characterized by a socioeconomic index of four [70]. Various public buildings are located to the northeast of Katznelson-Garden, including community centers, kindergartens, schools, and a synagogue. The park is surrounded by three-story residential buildings on the west, east, and south sides, with a parking lot to its north. Katznelson-Garden spans a total area of 2800 square-meters and is monitored by three city cameras covering a significant portion of the park (Figure 1). Katznelson-Garden features four access points. There are two access points from the northern side, one of which is a ramp designed to accommodate wheelchairs, strollers, and bicycles. The other northern access point consists of stairs. On the east side, there are two access points: one from a public building, which is often locked, and another that connects the park to other residential buildings and a kindergarten. Katznelson-Garden is designed as a linear park. We divided the park into several activity zones, as shown in Table 1 and Figure 3. The park included three seating areas (Nos. 1–3), seven path zones (Nos. 15–21), eight playground areas (Nos. 23–30), one lawn area (No. 9), three amenities (Nos. 10, 11, and 14), and six greenery areas (Nos. 6, 7, 8, and 22). It is important to mention that these areas include only the activity zone of the park monitored by CCTV cameras rather than the entire park. The playground area is covered with a man-made shade structure. Trees and greenery line the paths, while the park’s center features several trees with large canopies that provide shade. The seating areas, made of iron or wood, are well-maintained, as are the litter bins and drinking fountains. The lawn area is poorly maintained and lacks shade.
Remez-Garden is situated in the Skia neighborhood on the northwest side of Or-Yehuda, with a socioeconomic index of four [70]. The Skia neighborhood primarily consists of two-family detached houses and lacks parks, making Remez-Garden one of the few available in the area. The park is bordered by one-way roads on the south, two-way roads on the west, residential buildings on the north, and a 1.5 m-high concrete wall on the east. It spans 4000 square-meters and is shaped like a trapezoid. Remez-Garden has three CCTV cameras (Figure 2) covering various areas, including seating areas, paths, lawns, and playgrounds. Compared to Katznelson-Garden, a unique feature is a lawn area with dog play equipment. The park has three access points: one on the west and north connected to the neighborhood streets and another on the north, leading to a path that connects to four kindergartens. The park is divided into several activity zones, as shown in Table 1 and Figure 4, which include four seating areas (Nos. 1, 2, 3, and 14), five path zones (Nos. 9–13), five playground areas with various facilities (Nos. 21–25), two lawn areas (Nos. 4 and 5), two amenities (Nos. 6 and 8), and six greenery areas (Nos. 15–20). The playground area is covered with a man-made shade structure, whereas the other areas remain unshaded. The lawn area is maintained, but lacks shade. It is worth noting that these areas do not include all the areas of the park—only the areas captured by the CCTV camera.

3.2. Data Source

This study employed city surveillance cameras in the Or-Yehuda municipality, which were installed in urban parks for security monitoring—a practice prevalent in numerous cities globally. Although the cameras were initially installed for security purposes, early observations conducted by the researchers confirmed that the three cameras in each park effectively cover the majority of the active zones where visitors tend to spend the most time. These city cameras operate through CCTV, a system that transmits signals to a designated location, viewable on a restricted number of monitors. These monitors are situated in the municipal building of Or-Yehuda, where trained supervisors oversee the video footage. The data obtained from the CCTV are advantageous for assembling a large participant sample and systematically gathering information. CCTV video represents a structured dataset comprised of consistent video frames, making it amenable to analysis using deep learning techniques. The cameras in Or-Yehuda are typically positioned three meters above ground level, providing sufficient coverage to employ deep learning methods for extracting demographic information. The specific camera model used is the Hikvision 2MP full HD (Or Yehuda, Israel), offering a resolution of 1920 × 1080 at 24 frames per second.
The selected times span a variety of days within the week and hours throughout the day (Table 2). Data were collected from three cameras covering the park over a span of three days, totaling 36 h of video footage from each park. Specifically, the days chosen were Sunday, Tuesday, and Saturday, representing both weekdays and weekends (notably, Sunday is a standard working day in Israel). This selection provides a comprehensive dataset encompassing the week’s beginning, middle, and end. The times of day were categorized as follows: morning (8:00 a.m. to 11:59 a.m.), afternoon (12:00 p.m. to 3:59 p.m.), and evening (4:00 p.m. to 8:00 p.m.). The selected dates, 17, 19, and 23 April 2022, fell within the Passover holiday period. Our primary objective was not to capture a representative sample of park usage, but rather to concentrate on periods of peak activity. This approach was chosen to effectively demonstrate our methodology for analyzing gender differences in park utilization. By targeting periods of high activity volumes, we were able to effectively highlight and analyze gender-specific trends in park use. The selected dates in April were chosen strategically to align with this objective, as they provided optimal conditions for observing peak park attendance. The choice of days was guided by the assumption that holiday periods and warm, pleasant weather conditions would attract a higher number of visitors. The analysis focused on the time frame from 8:00 a.m. to 8:00 p.m., as these hours were observed to be the most active in the two parks, while activity levels dropped significantly after 8:00 p.m. due to darkness. Since our objective was not to evaluate typical daily park behavior, but to examine how visitors interacted with various park elements, this specific time frame allowed us to effectively demonstrate our methodology for analyzing gender-based park usage. The weather on these dates was predominantly sunny, thus minimizing potential weather-related disturbances in the observational data. The seasonal context was spring, with daytime temperatures ranging from 23 to 37 degrees Celsius and nighttime temperatures from 14 to 18 degrees Celsius.

3.3. Data Analysis

This study integrates computer vision and deep learning with geospatial analysis to examine usage patterns based on gender, as illustrated in Figure 5. This study leverages city surveillance cameras in urban parks, employing computer vision and deep learning techniques to analyze the data. Through this approach, key variables were identified, including visitor gender, visitation hours, and activity zones. These variables were subsequently processed using geospatial analysis, enabling the creation of a comprehensive mapping of gender-based usage patterns across the two selected case studies.
Two primary analytical techniques were utilized: geospatial analysis and computer vision/deep learning. Geospatial analysis was performed using ArcGIS to examine the spatial characteristics of the site area. Concurrently, video data extracted from city cameras in the two parks were analyzed using computer vision and deep learning techniques. This analysis produced detailed information on each park, identifying the appearances of individuals captured in the video frames by gender, surrounding physical features, days of the week, and hours of the day. Subsequently, a comprehensive spatial and quantitative analysis was conducted based on the processed video data.
The deep learning approach was chosen for its capability to efficiently analyze extensive video datasets, incorporating spatial and temporal information. This method has demonstrated success in applications like person retrieval and search, excelling in addressing challenges such as motion, diverse lighting conditions, varying camera perspectives, and object size differences. By training artificial neural networks to identify patterns in data, deep learning enables tasks like image classification and object detection [71]. However, due to the limited availability of models capable of accurately detecting gender within our dataset, we utilized two deep learning models to enhance the accuracy of our dataset. In the field of urban planning, and particularly in urban park planning, it is highly important to have a deep understanding of the number of users in the urban park area, where they are primarily located, their proximity to various facilities, and which facilities are more actively used in order to enable optimal spatial planning. The ability to transfer this knowledge to the spatial dimension of GIS holds significant importance from a planning perspective.
We focused on the systematic analysis of surveillance footage to conduct spatial analysis within urban parks. The process began with annotation of the data and training of deep neural networks for gender classification and pedestrian detection. This study utilized YOLOv4, an algorithm known for its real-time object detection capabilities [72]. Its efficiency in handling multiple objects simultaneously, along with its ability to provide high accuracy in object identification, made it an ideal choice for this study. We did not customize or modify this tool for our specific requirements. The entire process was performed in real-time with a single pass over an image, detecting multiple objects simultaneously. We selected YOLO for human detection due to its ease of use and growing support for a wide range of object classes.
For gender classification, this study employed ResNet-18, a deep Convolutional Neural Network (CNN) with 18 layers [73]. ResNets, or Residual Networks, are based on the concept of residual learning rather than direct learning. Essentially, these networks focus on learning the difference between the input data and the desired output. This is facilitated by skip connections, which enable data to bypass certain layers and move directly forward. These connections help preserve the magnitude of the correction signals during backpropagation, avoiding degradation that can occur when passing through multiple layers. ResNets are composed of stacked residual blocks, each containing a few layers and a skip connection. This architecture allows ResNets to become very deep, enabling them to tackle highly complex classification tasks. By simplifying the learning process, residual learning improves training efficiency and increases success rates.
The network was trained on an extensive dataset comprising over 100,000 partial images of pedestrians and was rigorously evaluated using approximately 4000 different images. The evaluation process reserved these 4000 different images exclusively for testing, ensuring it remained separate from the training phase. We achieved an 87% success rate in gender identification. The model successfully classified approximately 3400 out of 4000 tested images accurately. The evaluation process was rigorously designed, utilizing a reserved dataset of 4000 images exclusively for testing, ensuring that this dataset was not used during the training phase to maintain the integrity of the results. The training phase involved preparing data for the Neural Network, sourced from both publicly available databases and our dataset. Since our dataset was not annotated, we used the Computer Vision Annotation Tool (CVAT) to manually annotate the data. This included drawing bounding boxes around individuals and assigning gender labels based on human judgment. After training, the model underwent an inference stage using the reserved test dataset, achieving the reported accuracy.
To create an automated tool for image classification without human involvement, we trained an EfficientNet-based Neural Network [74]. This network is designed to analyze partial images of individuals and determine their gender. The training process involves iteratively passing examples through the network, which predicts a class for each image. These predictions are then compared to the actual classes, and the network’s layers are adjusted using gradient backpropagation to reduce errors. Key parameters in this training include the learning rate (controlling the pace of learning), the batch size (defining the size of data groups for processing), and the number of epochs (indicating the total training cycles). The network’s architecture remained unchanged, while additional parameters were focused on regularization techniques to avoid overfitting. These techniques included applying dropout and augmenting images during training.
Temporal data were meticulously extracted from the video footage using template matching techniques, leveraging consistent templates across camera outputs to ensure the precise timestamping of activities. The spatial analysis involved dividing pixel locations from the video frame into activity zones using manual segmentation maps. Based on the segmentation of each video frame, the identification of people, along with their gender and the timestamp, was converted into a CSV file. Then, we mapped the segmented park activity zones from the video frame with a coordinate system. Each camera in each park had a dedicated segmentation map that assigned specific pixels to a unique index. Simultaneously, GIS data for the corresponding areas were mapped to the same indices. This setup enabled us to accurately translate pixel positions from images into real-world coordinates. The process involved aligning the camera frames with visible park features in the footage. These features were segmented into pixel-based areas, which were then linked to a coordinate system. By establishing this link, we could integrate pixel data with spatial location information. Our dataset was constructed by detecting the appearances of individuals within the camera frames. However, the algorithm did not track individuals across frames. If a person moved from one part of the frame to another, the algorithm classified them as a new appearance. For this reason, we refer to these detections as “appearances” rather than “individuals.” The spatial classification analysis was conducted utilizing ArcGIS Pro 3.1.3 software. The Jenks Natural Break classification method was employed to evaluate the video analysis results. Through this method, we systematically grouped the numerical values into distinct categories, allowing us to discern and examine patterns by comparing the two parks.

4. Results

Katznelson-Garden recorded a total of 32,782 appearances, significantly higher than the 7992 appearances in Remez-Garden. The distribution of user appearances across different days of the week shows that Katznelson-Garden sees substantial activity, particularly on weekends. Specifically, there were 8888 appearances on Sunday, 5738 on Tuesday, and a notable peak of 18,076 on Saturday. Similarly, Remez-Garden showed increased usage on weekends, with 1476 appearances on Sunday, 891 on Tuesday, and 5625 on Saturday (Table 3). These highlight that both parks are more frequented during weekends.
When examining the findings by hours of the day, Katznelson-Garden had 3544 appearances in the morning, 12,194 in the afternoon, and 17,044 in the evening. Remez-Garden showed the same pattern with a smaller number of appearances. It recorded 1042 appearances in the morning, 1880 in the afternoon, and 5070 in the evening (Table 4). These results indicate that both parks experience peak usage in the evening.
Gender distribution across both parks reveals a substantial predominance of female users. In Katznelson-Garden, 81% of the total appearances were female, while male appearances were 19%. Remez-Garden exhibited a similar pattern, with female appearances at 86% and male appearances at 14%. This marked gender disparity is consistent across both parks. Further analysis of the hourly distribution by gender shows that in Katznelson-Garden, females were most prevalent during the evening (83%), followed by the afternoon (79%) and morning (76%). Conversely, males had the highest presence in the morning (24%), with a decreasing trend through the afternoon (21%) and evening (17%). A similar pattern is observed in Remez-Garden, where female appearances were 80% in the morning and 87% in the afternoon and evening. Male appearances followed a consistently lower distribution, with 20% in the morning and dropping to 13% in the afternoon and evening.
Katznelson-Garden is significantly more popular, with more user appearances across all hours of the day and days of the week. Both parks experience increased usage on weekends, with a pronounced peak in Katznelson-Garden on Saturdays. The gender disparity is evident in both parks, with female users predominantly outnumbering male users.

4.1. Distribution of Activity Patterns by Gender Across the Parks

In this section, we analyze the users’ appearances by park elements (such as seating areas, paths, playground facilities, etc.) and by gender (Table 5 and Table 6; Figure 6). The findings reveal distinct usage patterns in different activity zones.

4.1.1. Paths and Staying Areas

In both parks, paths exhibited high levels of activity, particularly by males. The path segment (No. 21) in Katznelson-Garden and the path segment (No. 9) in Remez-Garden are the most frequented path segments (Figure 6). These path segments differ in their shape, size, and functions. Katznelson-Garden’s path segment (No. 21), approximately 4.5 m in width, is a straight path shaped like a square, facilitating walking, running, and cycling. Path segment (No. 21) exhibited the highest usage, with 20.4% of users being male compared to 17.6% female (Table 5). Remez-Garden’s path segment (No. 9) is approximately 12 m wide and circle-shaped, with three access points to other path segments. Due to its shape and location, this space provides ample room for recreational activities, including children playing, engaging in group activities, and riding bicycles. This path segment also exhibited the highest usage, with 44.7% male and 42.7% female (Table 5). In general, most of the path segments in both parks are situated next to playgrounds and lawns.

4.1.2. Seating Areas

The seating area that showed the highest levels of activity was the bench (No. 1) in Katznelson-Garden, which observed a higher percentage of female users (9.3%) compared to male users (6.0%) (Table 5). In Remez-Garden, there were minimal gender differences in seating usage. The most active seating areas were those located near activity zones with high levels of activity. The bench (No. 1) in Katznelson-Garden is located in the playground area, whereas the picnic table (No. 14) in Remez-Garden is situated near the path (No. 9) (Figure 5). Both seating areas are characterized by their shaded locations.

4.1.3. Playground Facilities

The playground areas in both parks exhibited high levels of activity, particularly by females. In both parks, the most active areas were the synthetic grass area. In Katznelson-Garden, this was the playground area (No. 30), and in Remez-Garden, it was the playground area (No. 25) (Figure 5). The synthetic grass area (No. 30) was used by 20.4% of females compared to 12.9% of males. Similarly, the synthetic grass area (No. 25) was utilized by 5.7% of females compared to 2.9% of males (Table 5). Additionally, Remez-Garden’s concrete stage area (No. 21), located near the path segment (No. 9) and elevated above the path level, exhibited significant activity, with 8.3% of females and 6.1% of males (Table 5). Despite the absence of facilities on the stage, its elevation and location seem to encourage activity. The playground facilities in Katznelson-Garden were more intensively used than those in Remez-Garden. The swing area in Katznelson-Garden was utilized by 10.4% of females and 6.0% of males, whereas in Remez-Garden, the facilities were used by less than 0.6% of both genders (Table 5). The quantity of playground areas in the two parks differs, as does their centralization. In Katznelson-Garden, the playgrounds are distributed among small adjacent areas. In Remez-Garden, they are primarily concentrated around a single space. This spatial difference influences the activity patterns within them.

4.1.4. Lawns

Overall, the lawn areas exhibited different activity levels between the parks. In Katznelson-Garden, the lawn area (No. 9) demonstrated minimal usage, with only 0.2% of female and 0.3% of male users, indicating a low preference for lawn areas among both genders (Table 5). In Remez-Garden, the northwest lawn (No. 4) was used by 8.5% of females and 11.8% of males, and the lawn with dog facilities (No. 5) was used by 7.1% of females and 8.4% of males (Table 5). These differences suggest that males may be more likely to use lawn areas for activities such as dog walking. The lawn area (No. 4) is characterized by a lack of shade, where children and adults were observed walking and running. The lawn area with dog facilities (No. 5) features man-made shade, where children and adults were observed sitting, playing, and watching their dogs. These differences in usage underscore the variability in lawn usage based on factors such as activity type, size, and shade.

4.1.5. Amenities

Amenities, including litter bins and drinking fountains, showed low use in both parks. The most used amenities were the outdoor drinking fountain (No. 14) in Katznelson-Garden and the litter bins (No. 6) in Remez-Garden (Figure 5). In Katznelson-Garden, the outdoor drinking fountain (No. 14) was used by 1.3% of females and 1.7% of males, while the litter bins (Nos. 10 and 11) had equal usage of 0.4% for both genders. In Remez-Garden, the outdoor drinking fountain (No. 8) was used by 0.3% of females and 0.5% of males, and the litter bins (No. 6) had 1.5% of females and 1.2% of males (Table 5). These minor differences indicate that males may use drinking fountains slightly more frequently than females. Their usage remains low in both parks. This may indicate that there may be a need to relocate the amenities to enhance their usage.

4.1.6. Greenery

Greenery areas, such as flower beds and planters, were the least used elements in both parks. In both parks, the greenery elements were positioned along the edges of the park (Figure 5). In Katznelson-Garden, flower beds (Nos. 6–8) were less than 0.1% for both genders, similar to planters (No. 22). In Remez-Garden, planters (Nos. 16, 17, 18, and 19) also exhibited the lowest levels of usage (Table 5). The low utilization overall may be attributed to the design of the greenery, which serves primarily as a landscape setting rather than for activities.

4.1.7. Overall: Highest vs. Lowest Usage by Gender

When comparing the highest and lowest usage areas (Table 6), usage differences by element are evident. In Katznelson-Garden, playgrounds are highly utilized, with the south playground’s synthetic grass area being the most frequented at 18.93% and other playground areas also seeing substantial usage, indicating that playgrounds are major attractions. Conversely, in Remez-Garden, paths and staying areas dominate usage patterns, particularly the middle path at 42.94%. Lawns in Remez-Garden are highly used, especially the north-west lawn at 8.96%, reflecting a preference for open spaces where visitors can walk and engage in passive recreational activities. Overall, Katznelson-Garden emphasizes active, structured recreational activities with its playgrounds and paths, whereas Remez-Garden favors dynamic movement, open spaces, and staying areas.

4.2. Distribution of Activity Patterns by Gender Among Days of the Week Across the Two Parks

Three days were selected, each symbolizing a different part of the week: Sunday marks the start of the week, Tuesday signifies the middle, and Saturday denotes the end. It is significant to acknowledge that these periods are defined according to the Israeli calendar, in which Sunday is considered the first day of the week, and Saturday constitutes a part of the weekend.
Katznelson-Garden and Remez-Garden show distinct patterns of user distribution by gender and day of the week (Table 7). For Katznelson-Garden, the total number of appearances is 32,782, with 80% female and 20% male appearances. For Remez-Garden, the total number of appearances is 7992, with 86% female and 14% male appearances. The gender distribution highlights a significant disparity between female and male users in both parks. Katznelson-Garden exhibits a more balanced gender distribution compared to Remez-Garden, where female appearances are notably higher. This trend is consistent across all days of the week, with a particularly pronounced difference on Saturdays.
As shown in Figure 7, the spatial distribution demonstrates different patterns of the park elements across the days of the week. On Sundays, Katznelson-Garden shows a balanced distribution of males and females, with notable concentrations in the playground facilities (Nos. 23–30) and paths (Nos. 15–21). In contrast, Remez-Garden exhibits a higher density of users in its lawns (Nos. 4 and 5). Similarly, it exhibits a higher density in its playground facilities (Nos. 21–25) and its paths and staying areas (Nos. 9–13). On Tuesdays, both parks experience a drop in overall user appearances, but Katznelson-Garden maintains higher activity in its paths and staying areas (Nos. 15–21), while Remez-Garden sees moderate activity in its paths and staying areas (Nos. 9–13). Saturdays are the peak days for both parks, with Katznelson-Garden showing substantial usage in its playground facilities (Nos. 23–30) and paths (Nos. 15–21), especially in the northern and central areas of the park. Remez-Garden experiences high densities in its synthetic grass playground (No. 25) and picnic table seating areas (No. 14).
When comparing the usage between genders, there are distinct differences in the preferences for various park elements, which also vary across different days of the week. In Katznelson-Garden, female users show a marked preference for playground areas (Nos. 23–30) and paths and staying areas (Nos. 15–21) on Saturdays. On Tuesdays, the gender distribution remains similar, but the overall number of appearances is lower. In Remez-Garden, female users dominate the usage of playground areas and green spaces on Saturdays, with marked concentrations in the synthetic grass playground and planters, reflecting the overall higher activity levels. On Sundays, the pattern is similar, but the distribution of users is slightly different across elements. Both parks exhibit a pronounced gender disparity, with females significantly overtaking males, particularly in playground facilities and green areas. This trend is more marked in Remez-Garden, where female users dominate the usage of playground areas (Nos. 21–25), lawns (Nos. 4 and 5), and greenery (Nos. 15–20), indicating a higher preference for recreational facilities and green spaces. Katznelson-Garden shows a more balanced gender usage across its various elements, whereas Remez-Garden highlights specific zones with high female user density, particularly on weekends.
This finding suggests that both parks are more frequented during weekends. These results indicate a strong preference for park usage among female users, with a notable peak in weekend appearances.

4.3. Distribution of Activity Patterns by Gender Among Hours of the Day Across the Two Parks

The day is segmented into three time frames: morning (8:00–11:59 a.m.), afternoon (12:00–3:59 p.m.), and evening (4:00–8:00 p.m.). Overall, the results reveal distinct patterns in park usage between males and females across different hours of the day and between the Katznelson-Garden and Remez-Garden (Table 8).
Comparing the two parks by times of the day reveals that both parks show higher usage in the evenings. However, Remez-Garden has a higher percentage of total appearances in the evening (63%) compared to Katznelson-Garden (52%), with female dominance in appearances more pronounced in Remez-Garden (64%) compared to Katznelson-Garden (53%). Katznelson-Garden has a higher percentage of total appearances in the afternoon (37%) compared to Remez-Garden (24%), with a slightly higher percentage of male appearances in the afternoon in Katznelson-Garden. In the morning, Remez-Garden has a higher percentage of total appearances (13%) compared to Katznelson-Garden (11%), with a notably higher percentage of male appearances in Remez-Garden (19%) compared to Katznelson-Garden (13%).
As shown in Figure 8, the spatial distribution demonstrates different patterns of park element usage across various times of the day. In the morning, Katznelson-Garden shows lower activity levels with moderate use of paths (Nos. 15–21) and light use of playground facilities (Nos. 23–30) and seating areas (Nos. 1–3). By the afternoon, the garden experiences a noticeable increase in activity, particularly in the playground areas (Nos. 23–30) and paths (Nos. 15–21). In the evening, Katznelson-Garden reaches its peak usage, with substantial concentrations in the central and northern playground facilities (Nos. 23–30) and heavily occupied paths (Nos. 15–21). Similarly, Remez-Garden shows lower morning usage with moderate activity on paths (Nos. 9–13) and light use of playground facilities (Nos. 21–25). Afternoon usage increases, especially in the playground areas (Nos. 21–25) and lawns (Nos. 4 and 5), with higher activity levels observed in the evening, notably in the synthetic grass (No. 25) and the picnic table (No. 14).
When comparing the usage between females and males, distinct preferences for various park elements emerge, varying throughout the day. In Katznelson-Garden, female users are especially concentrated in the playground areas and paths during the evening, contributing to the overall higher user density in these elements. In Remez-Garden, female users dominate the playground and lawns during the evening, reflecting the overall higher activity levels. This trend is especially noticeable in Remez-Garden, where females show a higher preference for recreational and green spaces.
Overall, Katznelson-Garden shows a more balanced gender usage across its various elements, whereas Remez-Garden highlights specific zones with high female user density in the evening.

5. Discussions

5.1. Usage Patterns by Park Characteristics

This study serves as an initial application of the proposed methodological approach, applied to two case studies, and presents preliminary findings. These results are intended as a foundation for future research, highlighting the need for deeper insights through comparisons with a broader and more diverse sample of case studies.
Katznelson-Garden and Remez-Garden are small neighborhood parks, each approximately 2000 m2. They share similar physical environment envelopes and are located in neighborhoods with comparable socio-economic levels. Despite these similarities, we observed differing occupancy levels in these parks, both overall and across different days of the week. Katznelson-Garden recorded significantly higher numbers of people appearances than Remez-Garden, with nearly six times more appearances on Sunday and Tuesday and three times more on Saturday (Table 7). The differences in the overall occupation level between the parks may be influenced by the distance from the residential buildings to the park. In Katznelson-Garden, three sides of the park are bordered by residential buildings. In contrast, in Remez-Garden, only one side borders residential buildings. The windows of the buildings facing Katznelson-Garden function as “eyes on the street”, a concept described by Jacobs [51]. This phenomenon may enhance the perception of control over the surrounding environment and a sense of security [32].
Park safety is a key factor influencing park usage, shaped by elements such as adequate lighting, well-maintained pathways, and effective integration with surrounding neighborhoods. These features are especially significant in enhancing women’s perceptions of safety [67]. Both parks in this study demonstrated these characteristics, which likely contributed to the higher proportion of female visitors observed. Additionally, previous research underscores the role of well-maintained greenery in improving women’s sense of safety [68]. In this study, the lawn in Katznelson-Garden was observed to be poorly maintained, while the lawn in Remez-Garden was well-kept. This distinction aligns with our findings, as the green spaces in Remez-Garden attracted more female visitors compared to those in Katznelson-Garden, highlighting the impact of maintenance on park usage. These findings suggest that features shaping women’s perceptions of safety should be prioritized in urban design, serving as a foundation for developing guidelines in urban planning practices.
Both parks are designed with similar characteristics, including seating areas, paths and staying areas, lawns, playground facilities, amenities, and greenery. However, their occupancy levels differ, which may be attributed to their geometrical shape and layout. Katznelson-Garden is a linear park, while Remez-Garden is trapezoidal-shaped. The shape of the parks influences the layout of their features, potentially affecting their usage patterns and their use in this case [26]. The most active area in Remez-Garden is the central path and staying area, which is shaped like a half-circle and serves as the hub of the park’s walkway network. Similarly, in Katznelson-Garden, the path areas are the most active. Consistent with other studies, this study found that paths were the most active areas compared to other park activity zones [6]. In both parks, the paths are adjacent to the playground facilities, which are among the top active areas. Consistent with other studies, playground areas were found to be one of the most popular activity settings in parks [54].
Staying areas, such as playgrounds, lawns with dog facilities, and wide areas for path-and-stay areas, such as in Remez-Garden, provide designated spots for spending more time and engaging in diverse activities tailored to specific users’ characteristics. Meanwhile, passing areas can complement these zones for different age groups and genders. For example, studies have shown that elderly individuals prefer quieter places at a distance from playgrounds where they can still watch and enjoy children playing [75]. By strategically designing and placing these areas, urban parks can cater to varied needs and preferences, ensuring a harmonious and enjoyable environment for all visitors.
While the ten highest-ranked characteristics (Table 6) in Katznelson-Garden are concentrated within an area of 378 m2, the top ten usage occupations in Remez-Garden are spread over a larger area of 1196 m2, indicating a less compact distribution. The variation between compact and dispersed layouts significantly influences the differences in activity patterns observed in the parks. Compactness, as demonstrated by the concept of pocket parks, can significantly enhance a park’s usability [76]. In both parks, the playground area was identified as a significant source of attraction, serving as a hub with a high activity level. However, Remez-Garden offers additional amenities, including a lawn with dedicated dog facilities. Notably, Remez-Garden’s lawn with dog facilities was identified as one of the top three most popular attractions. This observation aligns with research suggesting that dog parks offer an activity that fulfills the needs of many urban residents [77]. These findings highlight the potential benefits of incorporating dog parks as designated activity zones within urban neighborhoods. Such spaces can provide valuable recreational opportunities for residents and should be considered as part of the planning guidelines for small neighborhood parks.
Both parks, despite their similarities in size and physical environment, exhibit differing occupancy levels due to their distinct geometrical shapes and layouts. Katznelson-Garden, being linear and closely bordered by residential buildings, benefits from a higher sense of security and control, resulting in higher overall occupancy rates compared to the trapezoidal-shaped Remez-Garden, which is not closely bordered by many residential buildings. The most active areas in both parks are the paths and staying areas adjacent to playground facilities, highlighting the importance of playgrounds as activity hubs. Additionally, Remez-Garden offers unique activities like lawns with dog facilities, which are highly popular. In both parks, the amenities experienced low usage. These findings suggest that designing urban parks with wide paths near playgrounds and compact layouts can enhance usability and attract more visitors.

5.2. Gender-Based Usage Patterns

The analysis of gender-based usage patterns in urban parks reveals significant differences in how men and women interact with and utilize urban parks. Both parks exhibit a marked predominance of female users, with females constituting 81% and 86% of the total appearances in Katznelson-Garden and Remez-Garden, respectively. In both parks, females predominantly frequent playground facilities, contributing to the overall higher user density in these areas. These findings align with other studies indicating that women predominantly occupy playground areas, usually involving childcare-related activities in shaded areas [59,66]. These patterns suggest that women’s outdoor leisure activities are largely centered on family and caregiving responsibilities, limiting their engagement in other recreational activities compared to men [61]. These findings emphasize the importance of designing neighborhood parks with diverse playground facilities, particularly in areas with families and young children. Such features can enhance the usability and appeal of parks, catering to the needs of local communities. While in both parks, men show a preference for paths, indicating a tendency towards more dynamic and movement-oriented activities, males exhibit higher morning activity, primarily utilizing paths for movement-based activities such as walking, cycling, and walking with dogs. These findings are consistent with other studies that found men more often engage in activities such as walking, cycling, and walking with dogs, often in less shaded areas [59,66].
In contrast, Remez-Garden displays a slightly different trend, with female users heavily occupying lawns with dog facilities and green spaces during the evening, reflecting a preference for recreational and green spaces. This preference may be due to the presence of fenced lawns with dog facilities and seating areas, which enhance the sense of safety. Studies suggest that women prioritize comfort and safety in public spaces [59,64]. The presence of men on the paths in the evening is consistent with other studies that found men prioritize vitality and entertainment, often seeking movement-oriented activities [64].
These gender-based usage patterns highlight the importance of considering specific park elements when designing urban parks to ensure they cater to different genders’ diverse needs and preferences. Men and women have different preferences and concepts of optimal public space experiences, necessitating thoughtful urban park design to accommodate these differences.

5.3. Gender-Based Usage Patterns by Time

The analysis of gender-based usage patterns in Katznelson-Garden and Remez-Garden reveals significant differences in how males and females utilize the parks throughout the day. In Katznelson-Garden and Remez-Garden, 53% and 64% of females show a marked preference for evening visits, respectively. This observation aligns with other studies [59,61,62,63]. Afternoon visits also highlight gender disparities. In Katznelson-Garden, 36% of female appearances occur in the afternoon compared to 40% for males, while in Remez-Garden, both genders account for 24% of the appearances, although males are slightly less frequent. In the morning, males are more prevalent in both parks; Katznelson-Garden reports 13% of male appearances in the morning compared to 10% for females, while Remez-Garden has 19% of male appearances versus 12% for females, as is also shown in other studies [59,61]. In contrast, other studies suggest that both females and males are more likely to visit parks in the morning [60,62]. These discrepancies might be due to the varying characteristics of the urban parks examined, suggesting a potential link between the time of day and gender-specific usage patterns influenced by different park features. These findings suggest that the design of urban parks should incorporate time-sensitive planning to cater to the differing needs of each gender. This can be achieved by engaging with local communities to understand gender-specific preferences and tailoring park activities accordingly. For example, morning activities might focus on male-oriented groups, such as running clubs, whereas evening programs might focus on playground activities designed to engage both children and their parents. These patterns underscore a notable gender difference, with females predominantly favoring evening and afternoon visits, especially in Remez-Garden, whereas males exhibit a relatively higher presence in the morning.

5.4. Limitation

This study detected 40,774 appearances (not the number of individuals) of visitors at two urban parks over 36 h. The number of appearances reflects activity levels at specific times and locations. The strength of this approach lies in the large sampling under investigation utilizing city cameras as a data source, encompassing spatial and temporal information along with user gender. This study examines data collected from three dates in April during a holiday period. These dates were strategically selected to align with the study’s objective of analyzing gender-specific patterns in park usage rather than to capture a representative sample of park usage. The analysis did not include data from workdays or other seasons with varying weather conditions. Future research could address these limitations by including data from workdays and examining park usage across a range of seasonal contexts. Additionally, the selected time frame of 8:00 a.m. to 8:00 p.m. over three days (36 h) concentrates on the most active periods identified in these case studies. This study does not include early morning activity (e.g., from 6:00 a.m.) or late-night activity (after 8:00 p.m.), which could offer insights into park usage during less active hours.
Several limitations are associated with using city cameras for data collection. Firstly, the municipality’s decisions regarding the location, angle, and number of cameras installed in the urban parks mean that certain areas remain unobserved, resulting in incomplete data for those regions. Secondly, as the cameras were primarily installed for security purposes, they were placed in potentially hazardous areas, leading to some regions being covered by multiple cameras. We excluded areas observed by multiple cameras to eliminate bias in our results.
Additionally, applying computer vision and deep learning methods for camera analysis presents several challenges. Blurriness in some image areas, particularly under low-light conditions, can impede accurate identification. The varying distances of people from the camera can cause significant differences in their size and appearance, complicating the identification process. To mitigate these issues, we employed high-resolution cameras strategically positioned to cover the maximum area, supplemented by lighting poles to ensure adequate illumination.

6. Conclusions

This study addresses the challenges of examining the relationship between users’ socio-demographic characteristics and urban park features, particularly when considering temporal and spatial dimensions with comprehensive data collection. By investigating the association between park characteristics and gendered usage patterns across different times of the day, this study provides new insights into urban park usage patterns. Through a comparative analysis of two urban parks, the study utilizes big data from CCTV cameras to explore human behavior. The methodology integrates traditional geospatial analysis with advanced computer vision techniques, offering an innovative framework for analyzing urban park usage. The findings provide urban planners and decision-makers with data-driven insights to support informed and evidence-based planning.
This research supports sustainable urban development goals by promoting inclusive and equitable public spaces, aligning with SDG 11: Sustainable Cities and Communities. By analyzing gendered and temporal usage patterns in urban parks, it provides valuable insights into designing parks that address the diverse needs of various demographic groups, thereby enhancing the quality of urban life. The study emphasizes the importance of gender-sensitive design, highlighting features such as playgrounds and shaded areas to promote equity.
Using a data-driven approach that combines computer vision and geospatial analysis, the research facilitates evidence-based urban planning and more efficient resource allocation. Its findings on compact versus dispersed park layouts offer practical strategies for optimizing land use by strategically designing spaces for specific activities. For instance, positioning activity zones in optimal locations—such as playgrounds near shaded paths for families or open lawns for dynamic uses—can significantly enhance park functionality and inclusivity. Compact layouts, exemplified by Katznelson-Garden, maximize usability by clustering high-activity zones within limited space. Conversely, dispersed layouts, such as Remez-Garden, provide diverse activity zones to cater to varied user preferences, but require careful spatial organization to ensure accessibility and sustained engagement.
In this study, we employ a comparative evaluation method to identify the park features associated with varying activity levels between genders in two urban parks. Traditional methods relying on self-reported data and manual observation are prone to biases and inaccuracies and often lack the capability to simultaneously integrate spatial–temporal dimensions and socio-demographic characteristics. This study utilized computer vision and deep learning techniques to identify people from video data. Using ArcGIS for spatial analysis and deep neural networks such as YOLOv4 and ResNet-18 to identify people and their gender, we systematically analyzed surveillance footage to derive detailed spatial and temporal information. The processed data were then mapped using the Jenks Natural Break classification method, allowing us to compare patterns of park usage by gender across the activity zone of the parks. This approach provides valuable insights into the influence of urban park features on user activity levels and gender.
The comparative analysis of Katznelson-Garden and Remez-Garden elucidates the significant influence of park layout on visitor activity patterns, particularly concerning the compactness of activity zones. Katznelson-Garden, characterized by a centralized and compact layout, exhibited a notably higher frequency of visitor appearances compared to Remez-Garden, which is characterized by a dispersed layout. Furthermore, the comparative evaluation between the parks revealed that both parks experienced peak usage in the evenings. Additionally, there were differences in the use of playground areas and paths, with varying frequencies of use observed across different genders. These findings underscore the importance of gender-sensitive urban design and highlight the influence of park features on user behavior.
This study provides new insights for urban planners and policymakers, underscoring the necessity for inclusive public open spaces that accommodate diverse needs. Additionally, the findings underscore the critical impacts on park planning and the strategic allocation of activity zones within parks. This includes considerations for the optimal placement of features in proximity to high- and low-usage areas. Such evidence-based planning is essential for advancing landscape architecture practices and ensuring that public open spaces effectively serve the needs of diverse demographic groups. Future research should integrate socio-demographic and spatial–temporal dimensions to enhance our understanding of urban park dynamics.

Author Contributions

S.G.-S.: Writing—original draft, Formal analysis, Conceptualization, Investigation, Visualization. O.B.: Data curation. D.S.-P.: Supervision, Conceptualization, Methodology, Validation, Writing—review & editing. P.P.: Supervision, Conceptualization, Methodology, Validation, Writing—review & editing. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

The study was approved by the Ethics Committee of the Technion—Israel Institute of Technology (protocol number 145270).

Informed Consent Statement

Because there is no identifiable participant data or direct interaction in this study, informed consent is not required according to the results of the institution’s review.

Data Availability Statement

The datasets presented in this article are not available because of legal restrictions.

Acknowledgments

The authors acknowledge the technical support of the Signal and Image Processing Lab (SIPL) at the Faculty of Electrical and Computer Engineering, Technion-IIT, Israel.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Inkinen, T.; Yigitcanlar, T.; Wilson, M. Smart Cities and Innovative Urban Technologies. Urban Technologies; Routledge: London, UK, 2021. [Google Scholar]
  2. Giffinger, R. Smart city: The importance of innovation and planning. In Smart Cities, Green Technologies and Intelligent Transport Systems: 8th International Conference, SMARTGREENS 2019, and 5th International Conference, VEHITS 2019, Heraklion, Crete, Greece, Revised Selected Papers; Springer International Publishing: Cham, Switzerland, 2021; pp. 28–39. [Google Scholar]
  3. Karvonen, A.; Cook, M.; Haarstad, H. Urban planning and the smart city: Projects, practices, and politics. Urban Plan. 2020, 5, 65–68. [Google Scholar] [CrossRef]
  4. Yigitcanlar, T. Smart cities: An effective urban development and management model? Aust. Plan. 2015, 52, 27–34. [Google Scholar] [CrossRef]
  5. Huang, H.; Yao, X.A.; Krisp, J.M.; Jiang, B. Analytics of location-based big data for smart cities: Opportunities, challenges, and future directions. Comput. Environ. Urban Syst. 2021, 90, 101712. [Google Scholar] [CrossRef]
  6. Huai, S.; Van de Voorde, T. Which environmental features contribute to positive and negative perceptions of urban parks? A cross-cultural comparison using online reviews and Natural Language Processing methods. Landsc. Urban Plan. 2022, 218, 104307. [Google Scholar] [CrossRef]
  7. Song, Y.; Wang, R.; Fernandez, J.; Li, D. Investigating sense of place of the Las Vegas Strip using online reviews and machine learning approaches. Landsc. Urban Plan. 2021, 205, 103956. [Google Scholar] [CrossRef]
  8. Volenec, Z.M.; Abraham, J.O.; Becker, A.D.; Dobson, A.P. Public parks and the pandemic: How park usage has been affected by COVID-19 policies. PLoS ONE 2021, 16, e0251799. [Google Scholar] [CrossRef]
  9. Wang, Z.; Zhu, Z.; Xu, M.; Qureshi, S. Fine-grained assessment of greenspace satisfaction at regional scale using content analysis of social media and machine learning. Sci. Total Environ. 2021, 776, 145908. [Google Scholar] [CrossRef] [PubMed]
  10. Angel, A.; Cohen, A.; Dalyot, S.; Plaut, P. Impact of COVID-19 policies on pedestrian traffic and walking patterns. Environ. Plan. B Urban Anal. City Sci. 2023, 50, 1178–1193. [Google Scholar] [CrossRef] [PubMed]
  11. Manley, E.J.; Orr, S.W.; Cheng, T. A heuristic model of bounded route choice in urban areas. Transp. Res. Part C Emerg. Technol. 2015, 56, 195–209. [Google Scholar] [CrossRef]
  12. Ji, H.; Qing, L.; Han, L.; Wang, Z.; Cheng, Y.; Peng, Y. A New Data-Enabled Intelligence Framework for Evaluating Urban Space Perception. SPRS Int. J. Geo-Inf. 2021, 10, 400. [Google Scholar] [CrossRef]
  13. Kang, Y.; Kim, J.; Park, J.; Lee, J. Assessment of perceived and physical walkability using street view images and deep learning technology. ISPRS Int. J. Geo-Inf. 2023, 12, 186. [Google Scholar] [CrossRef]
  14. Li, X.; Ratti, C. Mapping the spatio-temporal distribution of solar radiation within street canyons of Boston using Google Street View panoramas and building height model. Landsc. Urban Plan. 2019, 191, 103387. [Google Scholar] [CrossRef]
  15. Park, K. Park and neighborhood attributes associated with park use: An observational study using unmanned aerial vehicles. Environ. Behav. 2020, 52, 518–543. [Google Scholar] [CrossRef]
  16. Tay, L.; Jebb, A.T.; Woo, S.E. Video capture of human behaviors: Toward a big data approach. Curr. Opin. Behav. Sci. 2017, 18, 17–22. [Google Scholar] [CrossRef]
  17. Yamamoto, J.; Inoue, K.; Yoshioka, M. Investigation of customer behavior analysis based on top-view depth camera. In Proceedings of the 2017 IEEE Winter Applications of Computer Vision Workshops (WACVW), Santa Rosa, CA, USA, 24–31 March 2017; IEEE: New York, NY, USA, 2017; pp. 67–74. [Google Scholar]
  18. Antonakaki, P.; Kosmopoulos, D.; Perantonis, S.J. Detecting abnormal human behavior using multiple cameras. Signal Process. 2009, 89, 1723–1738. [Google Scholar] [CrossRef]
  19. Rezaei, B.; Christakis, Y.; Ho, B.; Thomas, K.; Erb, K.; Ostadabbas, S.; Patel, S. Target-specific action classification for automated assessment of human motor behavior from video. Sensors 2019, 19, 4266. [Google Scholar] [CrossRef] [PubMed]
  20. Ibrahim, M.R.; Haworth, J.; Cheng, T. Understanding cities with machine eyes: A review of deep computer vision in urban analytics. Cities 2020, 96, 102481. [Google Scholar] [CrossRef]
  21. Jung, C.; Awad, J.; Al Qassimi, N. Analyzing the users’ satisfaction levels and perceptions of the Dubai water canal for future waterfront development in UAE. Future Cities Environ. 2021, 7, 14. [Google Scholar] [CrossRef]
  22. Muleya, N.; Campbell, M. A multisensory approach to measure public space quality in the city of Bulawayo, Zimbabwe. Town Reg. Plan. 2020, 76, 56–71. [Google Scholar] [CrossRef]
  23. Giles-Corti, B.; Broomhall, M.H.; Knuiman, M.; Collins, C.; Douglas, K.; Ng, K.; Lange, A.; Donovan, R.J. Increasing walking. Am. J. Prev. Med. 2005, 28, 169–176. [Google Scholar] [CrossRef] [PubMed]
  24. Salama, A.; Remali, A.; MacLean, L. Deciphering urban life: A multi-layered investigation of St. Enoch square, Glasgow city centre. Archnet-IJAR Int. J. Archit. Res. 2017, 11, 137–156. [Google Scholar] [CrossRef]
  25. Edwards, N.; Hooper, P.; Trapp, G.; Bull, F.; Boruff, B.; Giles-Corti, B. Development of a Public Open Space Desktop Auditing Tool (POSDAT): A remote sensing approach. Appl. Geogr. 2013, 38, 22–30. [Google Scholar] [CrossRef]
  26. Goličnik, B.; Thompson, C.W. Emerging relationships between design and use of urban park spaces. Landsc. Urban Plan. 2010, 94, 38–53. [Google Scholar] [CrossRef]
  27. Marušić, B.G. Analysis of patterns of spatial occupancy in urban open space using behaviour maps and GIS. Urban Des. Int. 2011, 16, 36–50. [Google Scholar] [CrossRef]
  28. United Nations. Goal 11: Sustainable Cities and Communities 2023. United Nations Sustainable Development Goals. Available online: https://sdgs.un.org/goals/goal11 (accessed on 24 December 2024).
  29. Malchrowicz-Mosko, E.; Rozmiarek, M.; Poczta, J. Eco-sport in the space of modern city. Olimp. J. Olymp. Stud. 2021, 5, 128–140. [Google Scholar] [CrossRef]
  30. Marcus, C.C. Therapeutic landscapes. In Environmental Psychology and Human Well-Being; Elsevier: Amsterdam, Netherlands, 2018; pp. 387–413. [Google Scholar]
  31. Clough, N.L.; Vanderbeck, R.M. Managing Politics and Consumption in Business Improvement Districts: The Geographies of Political Activism on Burlington, Vermont’s Church Street. Urban Stud. 2006, 43, 2261–2284. [Google Scholar] [CrossRef]
  32. Gehl, J. Life Between Buildings: Using Public Space; Island Press: Washington, DC, USA, 2011. [Google Scholar]
  33. McAuliffe, C. Young people and the spatial politics of graffiti writing. In Identities and Subjectivities: Geographies of Children and Young People; Skelton, T., Ed.; Springer: Singapore, 2015; pp. 1–23. [Google Scholar]
  34. Rojo, L.M. Taking over the square: The role of linguistic practices in contesting urban spaces. J. Lang. Polit. 2014, 13, 623–652. [Google Scholar]
  35. Maruani, T.; Amit-Cohen, I. Open space planning models: A review of approaches and methods. Landsc. Urban Plan. 2007, 81, 1–13. [Google Scholar] [CrossRef]
  36. Campos-Sa’nchez, F.S.; Abarca-A’lvarez, F.J.; Reinoso-Bellido, R. Assessment of open spaces in inland medium-sized cities of eastern Andalusia (Spain) through complementary approaches: Spatial-configurational analysis and decision support. Eur. Plan. Stud. 2019, 27, 1270–1290. [Google Scholar] [CrossRef]
  37. Turna, N.; Bhandari, H. Role of parks as recreational spaces at neighborhood level in Indian cities. ECS Trans. 2022, 107, 8685. [Google Scholar] [CrossRef]
  38. Mu, B.; Liu, C.; Mu, T.; Xu, X.; Tian, G.; Zhang, Y.; Kim, G. Spatiotemporal fluctuations in urban park spatial vitality determined by on-site observation and behavior mapping: A case study of three parks in Zhengzhou City, China. Urban For. Urban Green. 2021, 64, 127246. [Google Scholar] [CrossRef]
  39. Nadzri, A.A.; Hussain, M.R.; Tukiman, I.; Al, M.; Zaini, A.M.; Shazalee, N.R. Physical and psychological health benefits of urban park in relation to pandemic crises. J. Soc. Sci. Humanit. 2023, 6, 16–21. [Google Scholar] [CrossRef]
  40. Xie, J.; Luo, S.; Furuya, K.; Sun, D. Urban parks as green buffers during the COVID-19 pandemic. Sustainability 2020, 12, 6751. [Google Scholar] [CrossRef]
  41. Sadeghian, M.M.; Vardanyan, Z. The benefits of urban parks, a review of urban research. J. Novel Appl. Sci. 2013, 2, 231–237. [Google Scholar]
  42. Sturm, R.; Cohen, D. Proximity to urban parks and mental health. J. Ment. Health Policy Econ. 2014, 17, 19–24. [Google Scholar]
  43. Li, H.; Ta, N.; Yu, B.; Wu, J. Are the accessibility and facility environment of parks associated with mental health? A comparative analysis based on residential areas and workplaces. Landsc. Urban Plan. 2023, 237, 104807. [Google Scholar] [CrossRef]
  44. Subiza-Perez, M.; Vozmediano, L.; San Juan, C. Green and blue settings as providers of mental health ecosystem services: Comparing urban beaches and parks and building a predictive model of psychological restoration. Landsc. Urban Plan. 2020, 204, 103926. [Google Scholar] [CrossRef]
  45. Ulrich, R.S.; Addoms, D.L. Psychological and recreational benefits of a residential park. J. Leis. Res. 1981, 13, 43–65. [Google Scholar] [CrossRef]
  46. Kajosaari, A.; Laatikainen, T.E. Adults’ leisure-time physical activity and the neighborhood built environment: A contextual perspective. Int. J. Health Geogr. 2020, 19, 35. [Google Scholar] [CrossRef]
  47. Liu, H.; Li, F.; Li, J.; Zhang, Y. The relationships between urban parks, residents’ physical activity, and mental health benefits: A case study from Beijing, China. J. Environ. Manag. 2017, 190, 223–230. [Google Scholar] [CrossRef] [PubMed]
  48. Cohen, D.A.; McKenzie, T.L.; Sehgal, A.; Williamson, S.; Golinelli, D.; Lurie, N. Contribution of public parks to physical activity. Am. J. Public Health 2007, 97, 509–514. [Google Scholar] [CrossRef] [PubMed]
  49. Baur, J.W.; Gómez, E.; Tynon, J.F. Urban nature parks and neighborhood social health in Portland, Oregon. J. Park Recreat. Adm. 2013, 31, 23. [Google Scholar]
  50. Alexander, C. A Pattern Language: Towns, Buildings, Construction; Oxford University Press: New York, NY, USA, 1977. [Google Scholar]
  51. Jacobs, J. The Death and Life of Great American Cities; Peregrine Books: London, UK, 1961. [Google Scholar]
  52. Lynch, K. The Image of the City; MIT Press: Cambridge, MA, USA, 1960. [Google Scholar]
  53. Whyte, W.H. The Social Life of Small Urban Spaces; Conservation Foundation: Washington, DC, USA, 1980. [Google Scholar]
  54. Baran, P.K.; Smith, W.R.; Moore, R.C.; Floyd, M.F.; Bocarro, J.N.; Cosco, N.G.; Danninger, T.M. Park use among youth and adults: Examination of individual, social, and urban form factors. Environ. Behav. 2014, 46, 768–800. [Google Scholar] [CrossRef]
  55. Kaczynski, A.T.; Besenyi, G.M.; Stanis SA, W.; Koohsari, M.J.; Oestman, K.B.; Bergstrom, R.; Reis, R.S. Are park proximity and park features related to park use and park-based physical activity among adults? Variations by multiple socio-demographic characteristics. Int. J. Behav. Nutr. Phys. Act. 2014, 11, 146. [Google Scholar] [CrossRef] [PubMed]
  56. McCormack, G.R.; Rock, M.; Toohey, A.M.; Hignell, D. Characteristics of urban parks associated with park use and physical activity: A review of qualitative research. Health Place 2010, 16, 712–726. [Google Scholar] [CrossRef] [PubMed]
  57. Arısoy, N. Examining the relationship between the socio-demographic characteristics of park visitors and park use by nonlinear canonical correlation analysis: The case study of Konya. J. Anatol. Environ. Anim. Sci. 2023, 8, 500–506. [Google Scholar] [CrossRef]
  58. Ma, Y.; Brindley, P.; Lange, E. The influence of socio-demographic factors on preference and park usage in Guangzhou, China. Land 2022, 11, 1219. [Google Scholar] [CrossRef]
  59. Huang, Y.; Napawan, N.C. “Separate but equal?” Understanding gender differences in urban park usage and its implications for gender-inclusive design. Landsc. J. 2021, 40, 1–16. [Google Scholar] [CrossRef]
  60. Krenichyn, K. Women and physical activity in an urban park: Enrichment and support through an ethic of care. J. Environ. Psychol. 2004, 24, 117–130. [Google Scholar] [CrossRef]
  61. Pérez-Tejera, F.; Valera, S.; Anguera, M.T. Using systematic observation and polar coordinates analysis to assess gender-based differences in park use in Barcelona. Front. Psychol. 2018, 9, 2546. [Google Scholar] [CrossRef] [PubMed]
  62. Scott, D. Exploring time patterns in people’s use of a metropolitan park district. Leis. Sci. 1997, 19, 159–174. [Google Scholar] [CrossRef]
  63. Van Hecke, L.; Van Cauwenberg, J.; Clarys, P.; Van Dyck, D.; Veitch, J.; Deforche, B. Active use of parks in Flanders (Belgium): An exploratory observational study. Int. J. Environ. Res. Public Health 2017, 14, 35. [Google Scholar] [CrossRef] [PubMed]
  64. Dhasmana, P.; Bansal, K.; Kaur, M. Assessing gender inclusive user preferences: A case of urban public spaces in Chandigarh. In Proceedings of the 2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT), Sakheer, Bahrain, 20–21 November 2022, Sakheer, Bahrain, 20–21 November 2022; IEEE: New York, NY, USA, 2022; pp. 221–226. [Google Scholar]
  65. Douglas, J.; Briones, M.; Bauer, E.; Trujillo, M.; Lopez, M.; Subica, A. Social and environmental determinants of physical activity in urban parks: Testing a neighborhood disorder model. Prev. Med. 2018, 109, 119–124. [Google Scholar] [CrossRef] [PubMed]
  66. Valera, S.; Casakin, H. Integrating observation and network analysis to identify patterns of use in the public space: A gender perspective. Front. Psychol. 2022, 13, 898809. [Google Scholar] [CrossRef]
  67. Braçe, O.; Garrido-Cumbrera, M.; Correa-Fernández, J. Gender differences in the perceptions of green spaces characteristics. Soc. Sci. Q. 2021, 102, 2640–2648. [Google Scholar] [CrossRef]
  68. Polko, P.; Kimic, K. Gender as a factor differentiating the perceptions of safety in urban parks. Ain Shams Eng. J. 2022, 13, 101608. [Google Scholar] [CrossRef]
  69. Israel Central Bureau of Statistics. Characterization and Classification Of Geographical Units by the Socio-Economic Level of the Population 2019. Available online: https://www.cbs.gov.il/en/mediarelease/Pages/2022/Characterization-and-Classification-of-Geographical-Units-by-the-Socio-Economic-Level-of-the-Population-2019.aspx (accessed on 3 February 2024).
  70. Israel Central Bureau of Statistics. Characterization and Classification of Statistical Areas Within Municipalities and Local Councils by the Socio-Economic Level of the Population 2015. Available online: https://www.cbs.gov.il/en/mediarelease/Pages/2019/Characterization-and-Classification-of-Statistical-Areas-Within-Municipalities-and-Local-Councils-by-the-Socio-Economic-Lev.aspx (accessed on 23 November 2023).
  71. Kong, Y.; Fu, Y. Human action recognition and prediction: A survey. Int. J. Comput. Vis. 2022, 130, 1366–1401. [Google Scholar] [CrossRef]
  72. Jiang, Z.; Zhao, L.; Li, S.; Jia, Y. Real-time object detection method based on improved YOLOv4-tiny. arXiv 2020, arXiv:2011.04244. [Google Scholar]
  73. He, K.; Zhang, X.; Ren, S.; Sun, J. Deep Residual Learning for Image Recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 27–30 June 2016; pp. 770–778. [Google Scholar]
  74. Tan, M.; Le, Q. EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks. Int. Conf. Mach. Learn. 2019, 97, 6105–6114, PMLR. [Google Scholar]
  75. Qiu, L.; Chen, Q.; Gao, T. The effects of urban natural environments on preference and self-reported psychological restoration of the elderly. Int. J. Environ. Res. Public Health 2021, 18, 509. [Google Scholar] [CrossRef] [PubMed]
  76. Zhang, H.; Kiyai, G. Pocket parks in urban design: Enhancing urban environment and community well-being. Highlights Art Des. 2024, 5, 7–12. [Google Scholar] [CrossRef]
  77. Urbanik, J.; Morgan, M. A tale of tails: The place of dog parks in the urban imaginary. Geoforum 2013, 44, 292–302. [Google Scholar] [CrossRef]
Figure 1. Katznelson-Garden’s areas captured by cameras.
Figure 1. Katznelson-Garden’s areas captured by cameras.
Sustainability 17 00865 g001
Figure 2. Remez-Garden’s areas captured by the cameras.
Figure 2. Remez-Garden’s areas captured by the cameras.
Sustainability 17 00865 g002
Figure 3. Map of Remez-Garden characteristics by activity zones.
Figure 3. Map of Remez-Garden characteristics by activity zones.
Sustainability 17 00865 g003
Figure 4. Map of Katznelson-Garden characteristics by activity zones.
Figure 4. Map of Katznelson-Garden characteristics by activity zones.
Sustainability 17 00865 g004
Figure 5. Research methodology framework.
Figure 5. Research methodology framework.
Sustainability 17 00865 g005
Figure 6. Comparison of spatial distribution of total appearances by gender between parks.
Figure 6. Comparison of spatial distribution of total appearances by gender between parks.
Sustainability 17 00865 g006
Figure 7. Comparison of spatial distribution of total appearances by gender and days between parks.
Figure 7. Comparison of spatial distribution of total appearances by gender and days between parks.
Sustainability 17 00865 g007
Figure 8. Comparison of the spatial distribution of total appearances by gender and time across the two parks.
Figure 8. Comparison of the spatial distribution of total appearances by gender and time across the two parks.
Sustainability 17 00865 g008
Table 1. Parks’ characteristics by activity zones.
Table 1. Parks’ characteristics by activity zones.
No.Katznelson-Garden CharacteristicsSize Area
(m2)
No.Remez-Garden CharacteristicsSize Area
(m2)
Seating1Bench seating area 61Bench seating area8
2Bench seating area (north lawn)72Bench seating area3
3Chair seating area (north)93Bench seating area 20
14Picnic table seating area 14
Paths15Path (east entrance) 33
16Path (middle)79
17Path (north entrance [ramp]) 712Path (north)100
18Path (north entrance [stairs]) 1910Path (north-mid-east)51
19Path (north playground) 2211Path (north mid)82
20Path (south lawn)3413Path (south mid)65
21Path (south playground)689Path (middle)254
Playground 23Playground—parallel bars (north-east)13
facilities24Playground—slides6525Playground—synthetic grass 282
25Playground—swing (north-east)623Playground—hopscotch game13
26Playground—swing (south-east)2622Playground—hedstrom9
27Playground—swing (south-mid)2924Playground—slides64
28Playground—swing (south-west)3121Playground—concrete stage area42
29Playground—synthetic grass (north)161
30Playground—synthetic grass (south)26
Lawns 5Lawn with dog facilities228
9Lawn2094Lawn (north-west)197
Amenities 10Litter bin16Litter bin2
11Litter bin28Outdoor drinking fountain1
14Outdoor drinking fountain2
Greenery6Flower bed (east)1515Planters (north-west)159
7Flower bed (northeast)1917Planters (north mid)80
8Planters (east)7816Planters (east lawn)30
22Planters (north entrance)8219Planters (south mid)171
20Planters (south-west)31
18Planters (south-east)80
Table 2. Selected time from video footage.
Table 2. Selected time from video footage.
Katznelson-GardenRemez-Garden
Number of days33
Number of hours per day1212
Total number of hours3636
Table 3. Distribution of total appearances by days of the week compared between parks.
Table 3. Distribution of total appearances by days of the week compared between parks.
Park SundayTuesdaySaturday
Katznelson-Garden Female82%71%83%
Male18%29%17%
Total100%
(n = 8888)
100%
(n = 5738)
100%
(n = 18,076)
Remez-GardenFemale81%82%88%
Male19%18%12%
Total100%
(n = 1476)
100%
(n = 891)
100%
(n = 5625)
Table 4. Distribution of total appearances by hours of the day compared between parks.
Table 4. Distribution of total appearances by hours of the day compared between parks.
Park Morning Afternoon Evening
Katznelson-GardenFemale76%79%83%
Male24%21%17%
Total100%
(n = 3544)
100%
(n = 12,194)
100%
(n = 17,044)
Remez-GardenFemale80%87%87%
Male20%13%13%
Total100%
(n = 1042)
100%
(n = 1880)
100%
(n = 5070)
Table 5. Distribution of users’ appearances by gender: a comparison between parks.
Table 5. Distribution of users’ appearances by gender: a comparison between parks.
Park ElementsNo.Katznelson-Garden CharacteristicsFemale (%)Male (%)No.Remez-Garden CharacteristicsFemale (%)Male (%)
Seating 3Chair seating area (north)1.2%1.9%1Concrete seating area (south)2.7%2.2%
2Bench (north lawn)0.2%0.2%14Picnic table 2.7%3.4%
1Bench 9.3%6.0%3Bench (south-east)0.0%0.0%
2Bench (middle)0.1%0.1%
Paths21Path (south playground)17.6%20.4%9Path (middle)42.7%44.7%
20Path (south lawn)1.7%3.9%7Path (north-west entrance) 0.6%0.4%
19Path (north playground) 5.3%7.9%13Path (south mid)2.5%2.0%
18Path (north entrance [stairs]) 4.8%6.0%10Path (north-mid-east)4.5%3.9%
17Path (north entrance [ramp]) 6.3%4.1%11Path (north mid)3.9%3.0%
16Path (middle)0.0%0.1%12Path (north)2.0%2.2%
15Path (east entrance) 8.0%10.1%26Path (south-west entrance) 4.1%4.7%
Playground facilities28Swing (south-west)0.1%0.0%21Concrete stage area8.3%6.1%
27Swing (south-mid)0.9%1.2%22Headstream0.6%0.4%
26Swing (south-east)10.4%6.0%24Slides0.4%0.4%
25Swing (north-east)0.1%0.0%23Hopscotch game0.2%0.1%
24Slides1.6%1.8%25Synthetic grass 5.7%2.9%
23Parallel bars (north-east)1.0%1.2%
30Synthetic grass (south)20.4%12.9%
29Synthetic grass (north)9.0%7.8%
Lawns9Lawn0.2%0.3%4Lawn (north-west)8.5%11.8%
5Lawn with dog facilities7.1%8.4%
Amenities 10Litter bin0.4%0.4%8Outdoor drinking fountain0.3%0.5%
14Outdoor drinking fountain1.3%1.7%6Litter bin1.5%1.2%
11Litter bin0.4%0.4%
Greenery7Flower bed (northeast)0.0%0.1%19Planters (south mid)0.1%0.7%
6Flower bed (east)0.1%0.1%18Planters (south-east)0.1%0.0%
22Planters (north entrance)0.1%0.1%15Planters (north-west)0.8%0.6%
17Planters (north mid)0.1%0.0%
20Planters (south-west)0.5%0.3%
16Planters (east lawn)0.0%0.1%
Table 6. Distribution of total appearances by the ten highest-ranked characteristics compared between parks.
Table 6. Distribution of total appearances by the ten highest-ranked characteristics compared between parks.
No.Urban Park CharacteristicsTotal (%)
HighestKatznelson-Garden30Playground—synthetic grass (south)18.93%
21Path (south playground)18.19%
26Playground—swing (south-east)9.53%
29Playground synthetic grass (north)8.78%
1Bench seating area8.69%
15Path (east entrance)8.37%
19Path (north playground)5.84%
18Path (north entrance [stairs])5.06%
17Path (north entrance [ramp])4.51%
12North entrance (ramp)2.22%
Remez-Garden9Path (middle)42.94%
4Lawn (north-west)8.96%
21Playground—concrete stage area8.02%
5Lawn with dog facilities7.27%
25Playground—synthetic grass5.31%
10Path (north-mid-east)4.43%
26South-west entrance4.19%
11Path (north mid)3.75%
14Picnic table seating area2.78%
1Bench seating area (south-east)2.67%
LowestKatznelson-Garden33Planters (north entrance)0.10%
22North entrance (stairs)0.08%
39Playground—swing (north-east)0.08%
42Playground—swing (south-west)0.06%
25Path (middle)0.05%
13Flower bed (northeast)0.04%
14Greenery (east)0.03%
18Litter bin0.03%
8East entrance0.03%
9Entrance to public buildings0.01%
Remez-Garden20Planters (south-west)0.46%
24Playground—slides0.43%
8Outdoor drinking fountain0.36%
19Planters (south mid)0.23%
23Playground—hopscotch game0.21%
18Planters (south-east)0.08%
17Planters (north mid)0.08%
2Bench seating area (middle)0.06%
16Planters (east lawn)0.05%
3Concrete seating area (south-east)0.03%
Table 7. Distribution of appearances by gender and days of the week across the parks.
Table 7. Distribution of appearances by gender and days of the week across the parks.
Park Female (%)Male (%)Total (%)
Katznelson-GardenSunday28%26%27%
Tuesday15%26%18%
Saturday57%48%55%
Total100%
(n = 26,392)
100%
(n = 6388)
100%
(N = 32,782)
Remez-GardenSunday17%26%18%
Tuesday11%15%11%
Saturday72%59%70%
Total100%
(n = 6894)
100%
(n = 1098)
100%
(N = 7992)
Table 8. Distribution of users’ appearances by gender and hours of the day across the two parks.
Table 8. Distribution of users’ appearances by gender and hours of the day across the two parks.
Park Female (%)Male (%)Total (%)
Katznelson-Garden Morning10%13%11%
Afternoon 36%40%37%
Evening53%47%52%
Total 100%
(n = 26,392)
100%
(n = 6388)
100%
(N = 32,782)
Remez-GardenMorning12%19%13%
Afternoon 24%22%24%
Evening64%59%63%
Total 100%
(n = 6894)
100%
(n = 1098)
100%
(N = 7992)
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Gravitz-Sela, S.; Shach-Pinsly, D.; Bryt, O.; Plaut, P. Leveraging City Cameras for Human Behavior Analysis in Urban Parks: A Smart City Perspective. Sustainability 2025, 17, 865. https://doi.org/10.3390/su17030865

AMA Style

Gravitz-Sela S, Shach-Pinsly D, Bryt O, Plaut P. Leveraging City Cameras for Human Behavior Analysis in Urban Parks: A Smart City Perspective. Sustainability. 2025; 17(3):865. https://doi.org/10.3390/su17030865

Chicago/Turabian Style

Gravitz-Sela, Shir, Dalit Shach-Pinsly, Ori Bryt, and Pnina Plaut. 2025. "Leveraging City Cameras for Human Behavior Analysis in Urban Parks: A Smart City Perspective" Sustainability 17, no. 3: 865. https://doi.org/10.3390/su17030865

APA Style

Gravitz-Sela, S., Shach-Pinsly, D., Bryt, O., & Plaut, P. (2025). Leveraging City Cameras for Human Behavior Analysis in Urban Parks: A Smart City Perspective. Sustainability, 17(3), 865. https://doi.org/10.3390/su17030865

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop