Urban Image Research Based On Social Media Data

The image of the city is not only an important way to explore the characteristics of the city,but also the core content of interpreting the public’s cognition of the city.It has important research value and significance in advocating the construction of humanism.However,due to the limitations of traditional social survey methods such as questionnaire surveys and on-site interviews in data acquisition and data depth,they have not been able to explore the intrinsic properties and true reflections of environmental imagery from the perspective of public cognition,resulting in a lack of in-depth exploration of the connotation of imagery in urban image research.The kinetic energy is separated from the public’s rich environmental perception and experience.With the rise of the Internet and mobile devices,social media has become an important communication medium for public cognition and expression.The non-intrusive and big-data content information represented by it has become an important resource for mining city image cognition.In such an opportunity Next,how to accurately grasp the scientific connotation of urban imagery,clarify the digital fitting relationship between urban imagery and social media data,establish a scientific and reasonable translation analysis method,and guide the corresponding urban planning and design is an important problem in the current urban imagery research..Based on the above problems and opportunities,the thesis tries to cut from the perspective of image translation of social media data,fully consider the overall characteristics of urban imagery and the theoretical connotation of the internal structure hierarchy,fit social media data corresponding to the attribute characteristics,and find data acquisition,processing,and processing Establish a multilevel and multi-dimensional urban image analysis method through a series of technological paths such as analysis,analysis,and visualization,and provide a basis for specific planning and design decisions while fitting the public perception and experience process through quantitative means.The thesis starts with the related theories of urban imagery,sorts out the essence of its various connotations,and concludes a set of connotation system suitable for the use of social media data to translate urban imagery,which is divided into two major levels,namely "overall characteristics" and "internal structure".As the directional core system of image exploration,it determines the object of specific exploration and analysis for the new media that introduces social media,and further discriminates the feasibility and necessity of the application of social media in the study of urban imagery.In terms of method models,from the perspective of digital fitting,the concept of "crowdsourcing" is introduced,which is intended to extract and translate the public spontaneous information represented by social media data in a structural way,and to introduce people who combine artificial intelligence and team interpretation.Computer collaborative technology methods to build the ability to process image data in a refined batch,so as to realize the mode of passive public participation in research and investigation,analyze the content of urban image in an efficient and nonintrusive way,and interpret the research of urban image in a new way Public participation.Finally,for the case city Wuhan,according to the logical sequence from shallow to deep,from information to system,two major method steps for using social media data to analyze the image of the city are proposed.On the one hand,construct the image connection between the high-identity,representative environmental consensus features and the data of Douyin and Xiaohongshu at the overall characteristic level,realize the image characteristic analysis of the urban scene impression and the strategy impression,and dig out the three circles of bright spots.The overall image characteristics of the level;on the other hand,the construction of the internal structure level consists of the image connection between the agglomeration and repetitive environmental organization elements and the contents of the microblog data to realize the four-dimensional level of the city’s "space-society-visual-emotion" Analysis of the image structure of.Finally,a multi-level collaborative multi-dimensional expansion of urban image analysis results has been formed,which overcomes the problems of operability,reliability and practicability of existing research methods,and expands the content depth and breadth of urban imagery.