地球信息科学学报 ›› 2020, Vol. 22 ›› Issue (5): 1033-1048.doi: 10.12082/dqxxkx.2020.190661

• “空间综合人文学与社会科学”专辑 • 上一篇    下一篇

城市创意产业空间动态集聚演化的计算与可视优化方法

周琦, 高长春*()   

  1. 东华大学旭日工商管理学院,上海 200051
  • 收稿日期:2019-11-05 修回日期:2020-03-17 出版日期:2020-05-25 发布日期:2020-07-25
  • 通讯作者: 高长春 E-mail:gcc369@dhu.edu.cn
  • 作者简介:周 琦(1983— ),男,博士生,主要从事城市创意产业组织管理、城市区域发展研究。E-mail:qiweekend@163.com
  • 基金资助:
    国家自然科学基金面上项目(71874027)

The Calculation and Visual Optimization Method of Spatial Dynamic Agglomeration Evolution of Urban Creative Industries

ZHOU Qi, GAO Changchun*()   

  1. Donghua University, Sunrise School of Business Administration, Shanghai 200051, China
  • Received:2019-11-05 Revised:2020-03-17 Online:2020-05-25 Published:2020-07-25
  • Contact: GAO Changchun E-mail:gcc369@dhu.edu.cn
  • Supported by:
    National Science Foundation of China(71874027)

摘要:

基于人文地理视角下的城市创意产业图像可视化分析对城市深层次空间综合和区域创新发展具有重大意义。但Swarm群智能动态时空建模难以满足创意产业空间集聚的可视化发展。本文研究目标是,从城市区域创意产业空间聚类影响因素指标出发,创新性地提出区域空间动态集聚轨迹算法(Density-Based Interest Spatial Clustering of Path,DBICP),并与计算机浏览器共建聚类可视化图像,为城市管理提供决策依据。首先,根据影响因素指标体系,利用2014—2018年空间卡口流量数据和产业指标数据进行预处理,构建空间标准聚类算法DBSCAN。然后,对其进行聚类密度分级优化形成全新DBICP算法并得出初步轨迹图像。最后,通过源码转译实现了浏览器界面下空间动态集聚轨迹图像的输出。结果表明:以上海市为例,普陀区、浦东新区、徐汇地区的创意产业空间分布形成了3种不同的聚类模式,并相应提出了分摊、均布、虹吸的管控策略。此方法克服了传统图像的聚类分级和轨迹测量的缺失,可以有效地从指标数据中发现图像轨迹聚类信息,体现了地理信息科学和人文社会学科的交叉融合。也为大数据动态图像的集聚方法提供了全新视角和借鉴价值。

关键词: 城市创意产业, 影响因素指标, 聚类算法, DBICP动态建模, 线性轨迹, 可视化图像, 上海市, 管控策略

Abstract:

Analysis of the urban creative industry' s image visualization based on the perspective of human geography is of great significance for the integration between urban deep space and regional innovation development. However, the intelligent dynamic spatiotemporal modeling of Swarm groups is insufficient in meeting the visual development of the spatial clustering of creative industries. This study aims to provide a basis for decision-making within city management. Starting from the influencing factors of spatial clustering of creative industries in urban areas, a novel process of density-based interest spatial clustering of path (DBICP) is proposed together with a computer browser to aggregate visual images. First, according to the indicator system of influencing factors, and through the space bayonet traffic data and industry indicator data during 2014 to 2018, preprocessing is performed for constructing a spatial standard clustering algorithm: The density-based spatial clustering of applications with noise. Second, a hierarchical optimization of clustering density is performed to develop a new DBICP algorithm and obtain a preliminary trajectory image. Finally, using source code translation, the output of spatially-aggregated trajectory images under the browser interface is completed. Through the selection of 7 creative spatial indicators, the selection of more than 4000 points of interest, 2 groups algorithm tests, data of 3 groups bubble-set preliminary planning, 3 sets of Canvas dynamic simulation sequential planning, and E-charts spatial dynamic partial planning are accomplished. The average moving trajectory distance is 4.88 km, the regional agglomeration degree is 0.84, and the dynamic agglomeration evaluation index is 5.01. The results of the process as applied to the sample city of Shanghai show that three different clustering patterns have been formed in the spatial distribution of creative industries in Putuo District, Pudong New District, and Xuhui District, thus evidencing the control response strategy of allocation, uniform distribution, and siphon. The vector clustering image generated by the method proposed in this paper can explore the clustering characteristics of smart dynamic activities of the urban big data in the future and can also effectively solve practical urban problems, such as business clustering graphical measurement and community traffic image survey, and provide relevant technical support and research means for the large-scale spatial dynamic clustering supervision of urban geography. The method overcomes the lack of clustering classification and trajectory measurement in traditional images, effectively finding clustering information of image trajectories from the index data. This in turn embodies the interdisciplinary integration of geographic and sociological information, thus providing a clustering method.

Key words: urban creative industry, indicators of influencing factors, clustering algorithm, DBICP dynamic modeling, linear trajectory, visualized image, Shanghai municipality, regulatory strategy