A Simplified Visual Analysis Method for Large-Scale Social Media Data

Abstract

The invention discloses a simplified visual analysis method for large-scale social media data. The method is as follows: using LDA model to analyze text topics of large-scale social media data, and project the results to low-dimensional space to intuitively present the topics distribution characteristics; quantify and count the distribution of different topic features, design an efficient sampling model to sample large-scale social media data, and maintain the topic feature distribution of the original data as much as possible while maintaining the spatial distribution of large-scale social media data; design Visualization and interactive analysis system tools, integrating topic analysis and sampling models, designing visual graphic interfaces such as topic river diagrams, histograms, matrix diagrams, and topic ring diagrams, and analyzing sampling results from different perspectives such as topic feature distribution, time series evolution, and spatial distribution. Evaluation, which enables users to perform simplified and exploratory analysis of large-scale geotagged social media data.

Description

technical field [0001] The invention belongs to the technical fields of news dissemination, graphics and visualization, and in particular relates to a simplified visual analysis method for large-scale social media data. Background technique [0002] The use and influence of social media in people's lives has far exceeded people's imagination. With the continuous development of the times and the continuous progress of society, social media has gradually become an indispensable part of people's lives. Li Xia et al. store and analyze social media data through Hadoop, thereby efficiently optimizing the index generation of the Solr search engine, and further analyzing social relationship grids, user groups, user emotions, customer city maps, and topic trends; Amir et al. A multi-method data analysis method was developed to obtain the spatio-temporal correlation between influenza-related data in social media data and actual influenza outbreak trends, so as to explain the behaviora...

AI Technical Summary

Problems solved by technology

[0004] However, due to the increasing scale of geotagged social media data, visual mapping elements overlap each other, which largely interferes with analysts' perception and evaluation of specific events happening in local areas

For example, when disasters occur, the actual spatial distribution of disaster-related social media datasets appears to be uniform, which means that analysts may not be able to effectively focus on key areas

Images