Urban Knowledge Graph Products and Their Applications
Cheng Weiqi
The Development of Smart Cities Requires New Tools
With the development of technologies such as the Internet of Things, cloud computing, and big data, the construction of smart cities has gradually improved from perceptual intelligence to cognitive intelligence. With the application of technologies like 5G, the urban perception capability has been further enhanced, resulting in faster, more comprehensive, and richer data collection. Urban data also includes text, images, audio, video, and other modalities, leading to increasing complexity in data. Currently, many cities have extracted data to urban big data centers, achieving partial intelligence, but the design of data structures still relies on traditional methods, causing the virtual space data to fail to map one-to-one with the real city’s operational data. To make good use of this data, it is necessary to organize it into a large knowledge base and use it as a foundational resource for smart cities.
What is an Urban Knowledge Graph?
To establish an urban knowledge base, a knowledge graph is essential. A knowledge graph is essentially a semantic network, a new generation of knowledge base technology that transforms information into knowledge through structured and semantic processing and applies it. The abstract representation of a knowledge graph consists of entities connected by semantics, transforming human understanding of the entity world into computer-understandable and computable semantic information in a structured manner. A knowledge graph can be understood as a networked knowledge base that reflects an entity and its related entities or events, with different entities interconnected through various attribute relationships, thus forming a network. Therefore, a knowledge graph can be seen as a symbolic representation of the physical world.
Traditional knowledge graph tools are mostly used in areas such as preference recommendation, knowledge Q&A, and user profiling, making it challenging to directly migrate to the complex information composite of a city. At the same time, urban operations also follow the inherent laws of urban development, so the premise for establishing an urban knowledge graph is to recognize and understand the city.
A city is a dense spatial regional socio-economic system. Within a city, various urban elements constrain and intertwine with each other, forming the complex systemic and holistic movement of the city. From the perspective of the city’s composition and structure, a city is a dynamic large system characterized by multiple dimensions, variables, levels, and factors, comprising both a complex system that integrates natural and artificial elements, and an ecological system with human society as the main body, with geographic space and various facilities as the environment. In this context, a city, as a large ecological system, encompasses social structure, artificial structure, resource structure, environmental structure, and even external structure. Only by achieving coordination of relationships within this large system and optimizing its structure can the overall functionality of the urban system be optimized.
An urban knowledge graph is a large knowledge base based on urban theory, covering urban planning, construction, management, and service fields. It provides a unified framework for sharing data based on governance agreements, utilizes open standards, and optimizes business and operations through the relationships between data. It uses common classification methods and vocabularies to describe the data of universally applicable terms across all departments in the city, linking various data from different sources on a comprehensive, queryable semantic graph to showcase relevant points in relationships.
The establishment of an urban knowledge graph is foundational work for urban data management and operational analysis. By semantically defining and instantiating urban data, it achieves standardized and knowledge-based management of different fields, industries, and types of urban data, thereby supporting complex urban analysis applications.
Shenzhou Holdings Urban Knowledge Graph Products and Applications
Shenzhou Holdings abstracts the city as an organic entity containing multiple layers, including natural landscape domains, physical infrastructure domains, population domains, cultural social governance domains, digital infrastructure domains, and ubiquitous technology domains. Each domain level in the city integrates through digitization and intelligence, providing efficient citizen services.
The urban knowledge graph product is a set of tools developed by Shenzhou Holdings based on a human-centered, layered structure urban theory, targeting business fields such as urban planning, design, construction, management, operation, and service. It organizes and manages massive urban data according to knowledge models, supporting urban knowledge computation and urban spatio-temporal analysis. Users can define urban knowledge models, combining knowledge extraction, knowledge fusion, and knowledge processing tools to achieve graph-based data storage and management. Additionally, it provides complex spatio-temporal analysis APIs, graph analysis APIs, and optimization algorithm APIs for urban business fields, such as path planning, urban scheduling, and smart enrollment applications, helping users quickly and cost-effectively build graph-based application scenarios. Furthermore, it also enables the visualization of urban data on maps through the provided visualization tools, showcasing knowledge display and exploration.
It relies on the aggregation and governance of urban data to construct knowledge network relationships between different urban data, revealing the intrinsic relationships of various urban elements from the perspective of overall urban operations, providing feedback adjustments based on the dynamic development trends of the city, thereby guiding urban operation management and decision analysis.
Here, an example in the educational scenario describes the application of urban knowledge graph products. The delineation of school districts and the assignment of age-appropriate children for enrollment are key concerns for administrative management departments. In the context of limited regional school resources and the dynamic distribution of age-appropriate students each year, coordinating the fair distribution of school resources and ensuring nearby enrollment is a significant challenge for managers. If traditional methods are used to achieve precise school district delineation, even with significant manpower and time investment, management departments may find their efforts futile.
By utilizing urban knowledge graph products, knowledge definitions of urban spaces (cities, counties, neighborhoods, school districts), urban objects (schools, teaching equipment, roads), urban natural persons (teachers, students), urban events (enrollment events), and urban time (enrollment time, belonging time period) can be established, forming a foundational knowledge model in the urban education field. This can be integrated with the CIM platform to establish urban entities based on CIM data, achieving mapping relationships between entity data and knowledge models, and storing data according to the graph model to create an urban education knowledge graph.
Once the graph construction is completed, visualization can be achieved through the product’s visualization tools, allowing urban graph information to be visually displayed based on maps. The image shows the corresponding relationship between community and school in a certain year and location, with the base map transparency polygon indicating the distribution of elementary school districts; the starting points of the lines represent various neighborhoods, while the endpoints are the target elementary schools; the starting point heat map indicates the distribution of students enrolling from each neighborhood.
In terms of intelligent analysis, spatial analysis APIs can be utilized to determine neighborhoods within the specified enrollment radius of each school; combined with the urban road traffic network, graph analysis APIs (shortest path) can be used to calculate the shortest enrollment distance for students, evaluating the enrollment situation of students from individual neighborhoods. When designing new student enrollments across the district, considering single and multiple school enrollments within the region, optimization algorithm APIs can be used to find the optimal enrollment distribution plan (with the shortest total enrollment distance) under the premise of meeting enrollment rules, serving as a reference for school district allocation. Additionally, statistical analysis tools can be employed to calculate the average teaching staff, teacher subject distribution, and equipment distribution for each school, guiding the optimal allocation of educational resources.
Conclusion
Shenzhou Holdings has developed urban graph products based on years of government data services and understanding of cities. Relying on the urban data provided by the CIM platform, combined with the dynamic operation, management, and service data of cities, it can quickly establish knowledge graphs across various industries, performing complex spatial analysis, graph analysis, and visualization functions, revealing the historical and current developments of cities, exploring the objective laws of urban operations, predicting future trends in urban development, and supporting the implementation of various applications in urban planning, construction, management, and services.
