Journal of Geo-information Science >
Analysis and Visualization of Multi-dimensional Characteristics of Network Public Opinion Situation and Sentiment: Taking COVID-19 Epidemic as an Example
Received date: 2020-05-28
Revised date: 2020-12-21
Online published: 2021-04-25
Supported by
National Natural Science Foundation of China(41871371)
National Key Research and Development Program of China(2016YFE0131600)
Copyright
At the beginning of 2020, COVID-19 epidemic swept across China, and the development of COVID-19 attracted extensive attention from all sectors of society. Social media platform is an important carrier of online public opinion. In the process of epidemic prevention and control, it is very important to analyze the characteristics of network public opinion comprehensively and accurately. Firstly, from the perspective of spatiotemporal correlation between public opinion ontology and object, we construct a multi-dimensional analysis model of network public opinion during the epidemic period. We obtained the network public opinion data related to the covid-19 epidemic in multiple media platforms from January 17 to March 17, 2020. Secondly, from the perspective of epidemic spread, the spatial and temporal evolution and semantic characteristics of network public opinion in Wuhan, Hubei and the national scale are explored by comparative study and Spearman correlation coefficient. Finally, we use HowNet sentiment dictionary and emotional vocabulary ontology to analyze public opinion sentiment, and use interactive information chart to visualize the above results. The results show that: (1) The characteristics of time changes of public opinions are basically the same in Wuhan, Hubei province and China. There is a positive correlation between the number of daily public opinions and the number of new cases per day. With the rapid spread of the epidemic, the number of daily public opinions continues to increase. As the epidemic is gradually brought under control, the number of daily public opinions has shown a tortuous downward trend. (2) There is a positive correlation between the spatial distribution of public opinion data and the distribution of epidemic situation. The spatial distribution of the number of public opinions is similar to the distribution of the epidemic situation, and the areas with a large number of public opinions are mostly areas with severe epidemics. Changes in public opinions are spatially related to the development of the epidemic. (3) During the epidemic, the neutral sentiment of online public opinions was the most. Compared with forums, WeChat and Weibo, news platforms have a more positive overall sentiment. (4) At different stages of the development of the epidemic, the emotional characteristics of Weibo hot search data are quite different. The mood changed from anxiety in the early stage of the epidemic to excitement in the mid-term. And as the epidemic is gradually brought under control, emotions have also stabilized. Generally speaking, there are more positive emotions than negative emotions. Research shows that the multi-dimensional analysis model proposed in this article can visually show the public opinions situation, public opinions focus, and emotional changes at multiple scales during the epidemic.
DU Yixian , XU Jiapeng , ZHONG Linying , HOU Yingxu , SHEN Jie . Analysis and Visualization of Multi-dimensional Characteristics of Network Public Opinion Situation and Sentiment: Taking COVID-19 Epidemic as an Example[J]. Journal of Geo-information Science, 2021 , 23(2) : 318 -330 . DOI: 10.12082/dqxxkx.2021.200268
图4 2020年1月17日—3月17日多尺度下的网络舆情数据累计数量分布注:该图基于自然资源部标准地图服务网站下载的审图号为GS(2019)1825号的标准地图制作,底图无修改。 Fig. 4 The distribution of network public opinions under multi-scale from January 17 to March 17, 2020 |
表1 2020年1月17日—3月17日不同媒体平台舆情情感倾向统计Tab.1 Statistics on sentiment trends of different media platforms from January 17 to March 17, 2020 |
媒体平台 | 正面舆情数据数量/条 | 中立舆情数据数量/条 | 负面舆情数据数量/条 | 正面舆情占比/% | 中立舆情占比/% | 负面舆情占比/% |
---|---|---|---|---|---|---|
微博 | 1199 | 9256 | 2194 | 9.48 | 73.18 | 17.35 |
微信 | 2544 | 19 152 | 3579 | 10.07 | 75.77 | 14.16 |
论坛 | 13 792 | 77 593 | 147 85 | 12.99 | 73.08 | 13.93 |
新闻 | 52 825 | 303 576 | 47 587 | 13.08 | 75.14 | 11.78 |
表2 2020年1月17日—3月17日正面及负面舆情数据数量各占比前五的地区Tab. 2 The top five regions in terms of positive public opinion and negative public opinion respectively from January 17 to March 17, 2020 |
正面舆情占比/% | 负面舆情占比/% | ||
---|---|---|---|
新疆 | 16.69 | 台湾 | 50.60 |
贵州 | 16.30 | 香港 | 25.30 |
江苏 | 14.91 | 江西 | 15.93 |
内蒙古 | 14.88 | 北京 | 14.51 |
河南 | 14.66 | 广东 | 14.30 |
[1] |
中华预防医学会新型冠状病毒肺炎防控专家组. 新型冠状病毒肺炎流行病学特征的最新认识[J]. 中华流行病学杂志, 2020(2):139-144.
[ Chinese medical association covid-19 prevention and control expert group. An update on the epidemiological characteristics of novel coronavirus pneumonia (COVID-19)[J]. Chinese Journal of Epidemiology, 2020(2):139-144. ]
|
[2] |
中华人民共和国中央人民政府. 习近平对新型冠状病毒感染的肺炎疫情作出重要指示[EB/OL]. http://www.gov.cn/xinwen/2020-01/20/content_5471057.htm, 2020-01-20.
[ The central people's government of the People's Republic of China. Xi jinping gave important instructions on the novel coronavirus pneumonia outbreak[EB/OL]. http://www.gov.cn/xinwen/2020-01/20/content_5471057.htm, 2020-01-20.]
|
[3] |
|
[4] |
|
[5] |
|
[6] |
|
[7] |
赖凯声, 付宏, 晏齐宏, 等. 地理舆情:大数据时代舆情研究的新路径[J]. 情报理论与实践, 2020,43(8):64-69.
[
|
[8] |
黄鑫楠. 热点信息关注度的时空特征及地理距离对其的影响作用[D]. 上海:华东师范大学, 2019.
[
|
[9] |
王卷乐, 张敏, 韩雪华, 等. COVID-19疫情防控中的中国公众舆情时空演变特征[J]. 地理学报, 2020,75(11):2490-2504.
[
|
[10] |
张琛, 马祥元, 周扬, 等. 基于用户情感变化的新冠疫情舆情演变分析[J]. 地球信息科学学报, 2021,23(2). DOI: 10.12082/dqxxkx.2020.200248. https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CAPJ&dbname=CAPJLAST&filename=DQXX20200924000&v=Bg3Ir%25mmd2BkXDeflqiQBFzhR5UXSvyr0P7LUFaG3ThKq1p2EBSB9a7iS5OblN0Cp7pm2.
[
|
[11] |
韩珂珂, 邢子瑶, 刘哲, 等. 重大公共卫生事件中的舆情分析方法研究——以新冠肺炎疫情为例[J]. 地球信息科学学报, 2021,23(2). DOI: 10.12082/dqxxkx.2020.200226. https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CAPJ&dbname=CAPJLAST&filename=DQXX20201116001&v=Bg3Ir%25mmd2BkXDeedZbQEQiBeLaVFYXRLYgI%25mmd2BQn7VOf%25mmd2BbBdY9EaOfjRPjkHHOGBazKfuJ.
[
|
[12] |
刘大均, 胡静, 程绍文, 等. 中国旅游微博空间分布格局及影响因素——以新浪旅游微博为例[J]. 地理科学, 2015,35(6):717-724.
[
|
[13] |
郑嘉丽, 张丰, 杜震洪, 等. 传染病的多尺度时空特征分析——以杭州市淋病、细菌性痢疾和流行性腮腺炎为例[J]. 浙江大学学报(理学版), 2018,45(5):605-616.
[
|
[14] |
|
[15] |
|
[16] |
|
[17] |
|
[18] |
|
[19] |
Douiji, yasmina, Mousannif, et al. Using YouTube comments for text-based emotion recognition[J]. Procedia Computer Science, 2016.
|
[20] |
|
[21] |
|
[22] |
Tencent news[EB/OL]. https://news.qq.com/zt2020/page/feiyan.htm.
|
[23] |
Baidu public opinion[EB/OL]. http://yuqing.baidu.com.
|
[24] |
Weibo[EB/OL]. https://s.weibo.com/top/summary/.
|
[25] |
王连喜. 网络舆情领域相关概念分布及其关系辨析[J]. 现代情报, 2019,39(6):132-141.
[
|
[26] |
高承实, 陈越, 荣星, 等. 网络舆情几个基本问题的探讨[J]. 情报杂志, 2011,30(11):52-56.
[
|
[27] |
刘凯, 秦耀辰. 论地理信息的尺度特性[J]. 地理与地理信息科学, 2010,26(2):1-5.
[
|
[28] |
李小文, 曹春香, 张颢. 尺度问题研究进展[J]. 遥感学报, 2009,13(s1):12-20.
[
|
[29] |
张昊旻, 石博莹, 刘栩宏. 基于权值算法的中文情感分析系统研究与实现[J]. 计算机应用研究, 2012,29(12):4571-4573,4597.
[
|
[30] |
陈建美. 中文情感词汇本体的构建及其应用[D]. 大连:大连理工大学, 2009.
[
|
[31] |
|
/
〈 |
|
〉 |