This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Medical Informatics, is properly cited. The complete bibliographic information, a link to the original publication on https://medinform.jmir.org/, as well as this copyright and license information must be included.
The COVID-19 pandemic is still undergoing complicated developments in Vietnam and around the world. There is a lot of information about the COVID-19 pandemic, especially on the internet where people can create and share information quickly. This can lead to an infodemic, which is a challenge every government might face in the fight against pandemics.
This study aims to understand public attention toward the pandemic (from December 2019 to November 2020) through 7 types of sources: Facebook, Instagram, YouTube, blogs, news sites, forums, and e-commerce sites.
We collected and analyzed nearly 38 million pieces of text data from the aforementioned sources via SocialHeat, a social listening (infoveillance) platform developed by YouNet Group. We described not only public attention volume trends, discussion sentiments, top sources, top posts that gained the most public attention, and hot keyword frequency but also hot keywords’ co-occurrence as visualized by the VOSviewer software tool.
In this study, we reached four main conclusions. First, based on changing discussion trends regarding the COVID-19 subject, 7 periods were identified based on events that can be aggregated into two pandemic waves in Vietnam. Second, community pages on Facebook were the source of the most engagement from the public. However, the sources with the highest average interaction efficiency per article were government sources. Third, people’s attitudes when discussing the pandemic have changed from negative to positive emotions. Fourth, the type of content that attracts the most interactions from people varies from time to time. Besides that, the issue-attention cycle theory occurred not only once but four times during the COVID-19 pandemic in Vietnam.
Our study shows that online resources can help the government quickly identify public attention to public health messages during times of crisis. We also determined the hot spots that most interested the public and public attention communication patterns, which can help the government get practical information to make more effective policy reactions to help prevent the spread of the pandemic.
The COVID-19 pandemic situation remains complicated, with nearly 82 million infection cases worldwide as of January 1, 2021 [
Pandemics are inherently negative situations; therefore, COVID-19–related news usually includes negative information such as infection rates, deaths, and quarantine information. Being surrounded by negative information can increase negative emotions, thereby driving perceptions of pandemic-related risk [
In this study, we analyze big data collected from popular online sources where people obtain, create, or discuss news and information in Vietnam, including Facebook, news websites, YouTube, forums, blogs, Instagram, and e-commerce sites. Data were collected from December 2019 to November 2020 to offer a wider view from diverse sources and a longer observation period. We analyzed this data to describe a pattern of the social reaction during two different waves of the COVID-19 pandemic in Vietnam using the issue-attention cycle and media framing theories as foundations to develop our research questions.
The risk of COVID-19 infection is still high worldwide given that vaccination is not yet widely used and some countries are trying to resume normal commercial operations, including commercial flights, in an attempt to recover economically from the consequences of the pandemic. The need to seek and discuss information during a pandemic crisis like COVID-19 is obvious. However, many people try to simplify complex information or rely on their current beliefs; this may create conflict if they must force new information into previous constructs. Facing the risk of illness or death, as in the COVID-19 pandemic, can change people’s attitudes toward “accepting information, handling and taking action on it” [
However, some researchers argued that the issue-attention cycle can differ depending upon culture [
Our study investigates public attention during the COVID-19 pandemic by examining internet discussion volume to find patterns and determine similarities or differences to the issue-attention cycle theory. The amount of public discussion on social media has changed over time based on the public’s response to each real event that occurred during the pandemic. Capturing the amount of public discussion not only helps to point out or compare patterns in issue-attention cycle theory but also shows how the public’s attention to specific events is different. From there, it is possible to help the government and stakeholders evaluate the severity of each event to the public and from there learn lessons for possible pandemic prevention in the future. Hence, the research questions related to this theory are:
RQ1: What is the level (volume) of public attention to COVID-19 in this study?
RQ2: What does the pattern of public attention to the pandemic look like?
Throughout the COVID-19 pandemic, a concurrent infodemic has bombarded the public, hindering the reception of reliable information sources so citizens can follow recommendations and protect themselves. Therefore, in addition to pointing out patterns of pandemic-related discussions, it is necessary to dig deep into sources that get the most public attention, which can help government and disease control centers stop inaccurate news that has reached a large number of people in a timely manner. These patterns can also help identify popular public channels to help legitimate agencies broadcast disease prevention messages more efficiently. Additionally, analyzing the public sentiment about the pandemic can help governments and the Centers for Disease Control and Prevention deliver more accurate prevention messages to appease public anxiety and insecurity. Therefore, we developed the third and fourth research questions:
RQ3: Which types of sources gained the most public attention and engagement during the pandemic?
RQ4: How did people react to the pandemic, as measured by expression of their emotions on social media?
The mechanism by which individuals create a clear conceptualization or reorient their thoughts about an issue is referred to as framing theory. The concept is based on the acceptance of an issue that can be presented from a number of viewpoints and is perceived as having implications for different principles or factors [
The explosive growth of information technology and social networking in the digital age has resulted in changes to the concept of “news,” which was once considered the product of a journalist [
RQ5: What frames are used and how frequently are they used in communications that occur during the pandemic? What main topics gained the most discussion and attention during the COVID-19 pandemic?
RQ6: Were different types of frames used during the first and second waves of the COVID-19 pandemic in Vietnam?
All information related to COVID-19 in Vietnam was obtained from the Ministry of Health of Vietnam’s official COVID-19 disease page [
This study aims to understand the public reaction to the COVID-19 pandemic via discussions among Vietnamese people on social media. We used SocialHeat, a fee-based social listening tool developed and sponsored by YouNet Group, to crawl data while following the terms of use from 7 types of sources: Facebook, Instagram, news, blogs, forums, e-commerce sites, and YouTube
The data set was collected from December 1, 2019, to November 13, 2020, from 63 million Facebook IDs (pages, individual profiles, and groups), 1.2 million YouTube accounts, 9000 news websites, and 300 forums in Vietnam. On account of the amount of data and technology limitations, we divided the timeline into 7 periods to crawl data, then reconnected the data in a complete and continuous timeline. To divide the timeline, we relied on highlighted events that took place during the period observed (December 1, 2019, to November 13, 2020). Specifically, we used data tracking new daily infections in Vietnam, which was updated by the Ministry of Health of Vietnam [
The 7 periods of the COVID-19 pandemic in Vietnam.
Period | Date | Total days, n | Events |
|
1 | Before January 23, 2020 | 54 | No confirmed cases in Vietnam | |
2 | January 23 to February 26, 2020 | 35 | First confirmed case in Vietnam; 16th infected case discharged from hospital | |
3 | February 27 to March 5, 2020 | 8 | No new cases in Vietnam | |
4 | March 6 to March 31, 2020 | 26 | 17th infected case confirmed and more reported afterward | |
5 | April 1 to April 15, 2020 | 15 | Implementation of social isolation | |
6 | April 16 to July 24, 2020 | 100 | No new cases in the community | |
7 | July 25 to November 13, 2020 | 112 | A new case in the community and the first deaths |
The volume of total mentions (a mention can be an original post, a comment, or a share) about COVID-19–related topics on digital channels, including Facebook, Instagram, news sites, forums, blogs, etc, was tallied and expressed by day to show how Vietnamese citizens reacted to COVID-19 pandemic–related events timeline-by-timeline. This study also integrates the real flow of facts and disease coping measures adopted by the government to analyze the relationship between government policies and peoples’ reactions during the pandemic.
To explore which sources attracted the most attention and engagement, we calculated the total interactions on COVID-19–related topics across all ID sources (Facebook, YouTube, and Instagram) and unique links on news, blog, forum, and e-commerce sites, then ranked them in order from highest to lowest. The total interaction with an engaging source equal to the total COVID-19–related posts was posted by observed source, plus total likes, shares, and comments that those posts gained.
Facebook is the most popular social platform in Vietnam [
We analyzed top posts created during the COVID-19 pandemic to understand which topics attracted the most citizen attention and their associated reactions via discussion sentiment analysis. Top posts were COVID-19–related posts that gained the most mentions (shares, comments) on Facebook, Instagram, YouTube, news sites, blogs, e-commerce sites, and forums.
Previous studies about the information shared on social media by users during crisis events had different ways of classifying content based on real events. For example, Vieweg [
The research on the nature of information spread about the COVID-19 pandemic on Weibo by Li et al [
For all COVID-19 data downloaded from Facebook, Instagram, news sites, forums, blogs, etc, the SocialHeat tool excluded noise, spam, and advertising posts before using natural language software developed by the YouNet Company for sentiment classification and to extract the top 50 keywords’ frequency for the 7 observed periods.
The most frequently mentioned keywords for each period were analyzed and visualized using VOSviewer (Nees Jan van Eck and Ludo Waltman) [
There was a total of 37,917,631 collectable mentions and 22,652,638 posts about COVID-19 from December 1, 2019, to November 13, 2020.
There was a positive correlation between total collectable mentions on social media and daily new COVID-19 infection cases (β0=74,451.4; β1=9366.9;
As shown in
During period 1 (December 1, 2019, to January 22, 2020), news sources gained the most public reaction, and TV channel sources followed right after. After period 1, community page sources steadily earned the most engagement. This was especially true during period 2 (January 23, 2020, to February 26, 2020), period 4 (March 6, 2020, to March 31, 2020), period 5 (April 1, 2020, to April 15, 2020), and period 7 (July 25, 2020, to November 13, 2020), when around 50% of Vietnamese citizens’ interactions about the pandemic came from community page sources. Meanwhile, TV channel sources (periods 1, 2, 4, and 7) and KOL sources (periods 3, 5, and 6) alternated second place status in terms of engagement on COVID-19–related topics.
In contrast, forum (periods 2, 6, and 7) and government (periods 1 and 3) sources gained less total interaction.
Total reactions on the top 500 most engaging sources.
Source | Period 1 (n=265,679), n (%) | Period 2 (n=3,217,036), n (%) | Period 3 (n=137,642), n (%) | Period 4 (n=39,200,553), n (%) | Period 5 (n=10,086,791), n (%) | Period 6 (n=567,114), n (%) | Period 7 (n=96,832,404), n (%) |
Community page | 63,251 (23.81) | 1,600,342 (49.75) | 50,549 (36.72) | 21,392,949 (54.57) | 4,734,704 (46.94) | 170,131 (30.00) | 42,131,669 (43.51) |
Forum | 5819 (2.19) | 16,918 (0.53) | 498 (0.36) | 540,775 (1.38) | 128,025 (1.27) | 0 (0.00) | 528 (0.00) |
Government | 346 (0.13) | 85,434 (2.66) | 0 (0.00) | 1,350,699 (3.45) | 431,201 (4.27) | 98,005 (17.28) | 8,340,435 (8.61) |
Group | 11,934 (4.49) | 266,640 (8.29) | 15,806 (11.48) | 3,367,923 (8.59) | 727,409 (7.21) | 38,046 (6.71) | 7,794,693 (8.05) |
Key opinion leaders | 39,234 (14.77) | 309,160 (9.61) | 45,529 (33.08) | 3,889,108 (9.92) | 1,903,164 (18.87) | 165,804 (29.24) | 11,732,842 (12.12) |
News | 73,982 (27.85) | 346,683 (10.78) | 15,043 (10.93) | 3,300,849 (8.42) | 724,832 (7.19) | 63,080 (11.12) | 10,681,121 (11.03) |
TV channel | 65,330 (24.59) | 359,969 (11.19) | 3863 (2.81) | 4,107,847 (10.48) | 1,384,393 (13.72) | 26,499 (4.67) | 16,057,180 (16.58) |
Unknown | 5783 (2.18) | 231,890 (7.21) | 6354 (4.62) | 1,250,403 (3.19) | 53,063 (0.53) | 5549 (0.98) | 93,936 (0.10) |
Total interactions on the most engaging sources were calculated by summarizing the number of each source’s COVID-19–related posts, likes, shares, and comments. We analyzed the average interaction on the top 500 most engaging sources to understand the efficiency of each COVID-19–related post created by each source.
As can be seen from
The average interaction on the top 500 most engaging sources.
Sources | Period 1 (n=10,167), n (%) | Period 2 (n=102,421), n (%) | Period 3 (n=3184), n (%) | Period 4 (n=1,111,044), n (%) | Period 5 (n=645,151), n (%) | Period 6 (n=12,802), n (%) | Period 7 (n=3,745,249), n (%) |
Community page | 427 (4.20) | 5443 (5.31) | 468 (14.70) | 74,540 (6.71) | 16,327 (2.53) | 915 (7.15) | 861,254 (23.00) |
Forum | 1940 (19.08) | 5639 (5.51) | 249 (7.82) | 90,129 (8.11) | 32,006 (4.96) | 0 (0.00) | 528 (0.01) |
Government | 346 (3.40) | 42,717 (41.71) | 0 (0.00) | 450,233 (40.52) | 431,201 (66.84) | 4900 (38.28) | 321,134 (8.57) |
Group | 385 (3.79) | 5442 (5.31) | 368 (11.56) | 57,083 (5.14) | 13,989 (2.17) | 865 (6.76) | 87,847 (2.35) |
Key opinion leaders | 162 (1.59) | 3964 (3.87) | 149 (4.68) | 54,015 (4.86) | 17,954 (2.78) | 825 (6.44) | 2,330,684 (62.23) |
News | 2000 (19.67) | 11,556 (11.28) | 1003 (31.50) | 94,310 (8.49) | 25,887 (4.01) | 2426 (18.95) | 67,245 (1.80) |
TV channel | 4666 (45.89) | 17,998 (17.57) | 644 (20.23) | 228,214 (20.54) | 81,435 (12.62) | 2409 (18.82) | 75,552 (2.02) |
Unknown | 241 (2.37) | 9662 (9.43) | 303 (9.52) | 62,520 (5.63) | 26,352 (4.08) | 462 (3.61) | 1005 (0.03) |
The type of COVID-19–related content that received the most attention varied from time to time. Starting from phase 2 onward, the diversity of content types increased to include caution and advice, policy reaction, and international situation updates (
Top posts with the most comments or shares.
Categories | Period 1 (n=49,480), n (%) | Period 2 (n=337,865), n (%) | Period 3 (n=24,173), n (%) | Period 4 (n=1,375,260), n (%) | Period 5 (n=312,851), n (%) | Period 6 (n=986,403), n (%) | Period 7 (n=9,161,011), n (%) |
Caution and advice | 27,607 (55.79) | 28,869 (8.54) | 0 (0.00) | 191,017 (13.89) | 28,067 (8.97) | 129,076 (13.09) | 347,096 (3.79) |
Notifications or measures have been taken | 7116 (14.38) | 13,222 (3.91) | 2229 (9.22) | 162,502 (11.82) | 0 (0.00) | 23,978 (2.43) | 2,120,104 (23.14) |
Donation money, goods, or services | 0 (0.00) | 0 (0.00) | 0 (0.00) | 12,167 (0.88) | 4982 (1.59) | 89,919 (9.12) | 1,477,020 (16.12) |
Emotional support | 0 (0.00) | 0 (0.00) | 4905 (20.29) | 328,056 (23.85) | 153,652 (49.11) | 195,632 (19.83) | 970,817 (10.60) |
Help seeking | 0 (0.00) | 0 (0.00) | 0 (0.00) | 9045 (0.66) | 0 (0.00) | 0 (0.00) | 133,675 (1.46) |
Doubt casting and criticizing | 0 (0.00) | 112,297 (33.24) | 3284 (13.59) | 186,581 (13.57) | 2537 (0.81) | 0 (0.00) | 666,394 (7.27) |
Counter rumors | 0 (0.00) | 55,397 (16.40) | 0 (0.00) | 0 (0.00) | 0 (0.00) | 0 (0.00) | 0 (0.00) |
Policy reaction | 1226 (2.48) | 116,869 (34.59) | 11,642 (48.16) | 289,200 (21.03) | 112,505 (35.96) | 0 (0.00) | 1,132,238 (12.36) |
International situation updating | 13,531 (27.35) | 11,211 (3.32) | 1678 (6.94) | 196,692 (14.30) | 11,108 (3.55) | 107,243 (10.87) | 100,369 (1.10) |
Medical issues: treatment, vaccine | 0 (0.00) | 0 (0.00) | 0 (0.00) | 0 (0.00) | 0 (0.00) | 432,389 (43.83) | 0 (0.00) |
Effects of the pandemic on the economy | 0 (0.00) | 0 (0.00) | 435 (1.80) | 0 (0.00) | 0 (0.00) | 8166 (0.83) | 0 (0.00) |
Entertainment | 0 (0.00) | 0 (0.00) | 0 (0.00) | 0 (0.00) | 0 (0.00) | 0 (0.00) | 2,213,298 (24.16) |
In period 1, when the first COVID-19 cases were found in Wuhan and had not yet spread to Vietnam, it is understandable that content concerning
After all text data related to COVID-19 was crawled, it was processed by a sentiment analysis tool developed by SocialHeat. All discussions were evaluated and sorted into one of three emotional categories, positive, negative, and neutral, based on natural language. In general, we found that people’s emotions when discussing COVID-19–related topics fluctuate and are unstable. Emotional neutrality almost always took first place. This is understandable because sources from government organizations and especially television and newspapers are expected to report “independent, reliable, accurate, and comprehensive information” [
Sentiment trend line from December 1, 2019, to November 13, 2020.
During period 1, when Vietnam had not yet recorded any cases and the pandemic situation had just begun in China, people learned about COVID-19 through information from the Ministry of Health and the press, so their mentality was still stable and optimistic. In period 2, when the first cases were discovered in Vietnam, people become more confused and worried. In period 3, negative emotions exploded when patient 17 was confirmed and there was a risk of community disease spread. Anger, blame, and anxiety were evident through the negative emotions expressed in the text lines discussed on social networks at that time. In period 7 when Vietnam experienced its second wave of COVID-19 with the re-emergence of community infection and the first recorded COVID-19 deaths, the optimism shown through positive emotions overwhelms the negative emotions expressed during this period. People have gradually adapted to the pandemic after experiencing the first wave and have confidence in the government’s ability to control the pandemic; positive signals that a Vietnamese COVID-19 vaccine would soon enter the human testing phase may have also contributed to the positive outlook [
The top 50 keywords were compiled and ranked in order from all discussions on the COVID-19 pandemic topic gathered during the study period. However, of the top 50, many keywords are synonyms, so we have grouped them into 36 keywords. The content of the top keywords was related to 4 main groups, including
Top 36 keywords for COVID-19–related topics during the COVID-19 pandemic in Vietnam from December 1, 2019, to November 13, 2020.
Rank | Word | Frequency, n |
1 | epidemic | 16,373,688 |
2 | COVID-19 | 10,720,319 |
3 | patient | 9,838,764 |
4 | quarantine | 9,783,246 |
5 | medical | 9,349,795 |
6 | go | 8,428,703 |
7 | hospital | 8,257,058 |
8 | Vietnam | 7,789,027 |
9 | case | 7,643,082 |
10 | disease | 6,397,632 |
11 | Danang | 5,621,518 |
12 | against | 5,208,937 |
13 | city | 5,066,441 |
14 | infected | 4,734,692 |
15 | virus | 4,099,245 |
16 | situation | 3,950,498 |
17 | information | 3,854,982 |
18 | province | 3,850,230 |
19 | prevention | 3,692,883 |
20 | citizen | 3,357,150 |
21 | mask | 3,316,868 |
22 | way | 3,310,041 |
23 | test | 2,784,690 |
24 | contact | 2,764,565 |
25 | government | 2,705,459 |
26 | Hanoi | 2,422,218 |
27 | family | 2,366,785 |
28 | money | 2,177,978 |
29 | coronavirus | 2,039,959 |
30 | The US | 1,903,865 |
31 | vehicle | 1,842,155 |
32 | result | 1,821,271 |
33 | treatment | 1,790,499 |
34 | Bach Mai hospital | 1,663,212 |
35 | together | 1,654,416 |
36 | get sick | 1,475,917 |
To better understand the context behind the most mentioned keywords and to highlight the top concerns about the COVID-19 pandemic expressed in internet discussions in each period in Vietnam, we extracted the top 50 keywords for each stage and visualized the associations between the keywords using VOSviewer software. The larger the dots, the more weight (frequency) that keyword possessed. The thicker and closer the link between two keywords, the more frequently both keywords appear.
The relationship between the top keywords in period 1 when no infections were found in Vietnam is shown in
Co-occurrences of the top keywords in period 1. SARS: severe acute respiratory syndrome.
In period 2 (
Period 3 (
Period 4 (
During period 5 (
Co-occurrences of the top keywords in period 2. WHO: World Health Organization.
Co-occurrences of the top keywords in period 3.
Co-occurrences of the top keywords in period 4.
Co-occurrences of the top keywords in period 5.
The most prominent keywords in period 6 (
The most prominent keyword in Vietnam during period 7 (
Co-occurrences of the top keywords in period 6. WHO: World Health Organization. VND: Vietnamese Dong.
Co-occurrences of the top keywords in period 7.
The COVID-19 pandemic is a sensitive time, and the need for reliable sources to avoid an
Listening to people’s attitudes during a pandemic as expressed through their interactions on the internet can help governments and related agencies quickly adjust communication plans to lead people through the pandemic with better precision. This study provides valuable information to those concerned about the COVID-19 pandemic in general and the public’s response to an entirely new crisis in particular. Based on the results of this study, governments could use it as a reference to evaluate the efficiency of using big data to address public health management issues. This resource not only can be used as a reference to deal with future epidemic crises but also is a valuable comparison of public reaction toward the pandemic across countries.
The volume of public attention during the COVID-19 pandemic was substantial, with a total of 37,917,631 public mentions and 22,652,638 public posts during the research observation period from December 1, 2019, to November 13, 2020. During the peak period, we recorded more than 1,255,175 publicly discussed mentions showing particular interest in the pandemic; these mentions also demonstrated that the amount of pandemic-related information generated by the public is substantial. This can inadvertently create an information matrix or
During the two COVID-19 pandemic waves in Vietnam from December 2019 to November 2020, the pattern of public attention looks similar to the issue-attention cycle described by Downs [
When the first COVID-19 cases were discovered in Wuhan, China, people were not too concerned about this strange disease, despite the attention given to it by the Vietnamese government, especially the Ministry of Health and related agencies. However, when Vietnam saw its first cases of infection, people began to pay more attention. Anxiety peaked when people became aware that this is a dangerous, contagious, potentially fatal disease and that there was no vaccine yet. The situation was eased when the government’s pandemic prevention responses were effective.
During the second COVID-19 pandemic wave in Vietnam, people remained interested in the pandemic but discussed it less on social networks. Public attention peaked with the first COVID-19–related deaths in Vietnam. Public attention then quickly dropped and diverted to other issues. This shows that although the second COVID-19 pandemic wave in Vietnam appeared to have a more negative factor (the first recorded deaths), the public’s attitude was not as intense as it had been during the first COVID-19 pandemic wave. This may be explained by people’s acceptance of the fact that death is a foreseeable outcome for patients infected with COVID-19 and at the same time an expression of
Per our data analysis, community pages on Facebook received the most total interaction from the public, likely because these aggregate information for the community with diverse content types. Each of these news sites usually post multiple articles per day on the same COVID-19 topic. However, in terms of average efficiency per article, government-controlled news sites outperformed other news sources. Drawing from this conclusion, we recommend that the government increase the number of articles posted to sources under its control to achieve the greatest dissemination of information to the community. In addition, the government can also coordinate with sources such as community pages and KOLs’ pages to quickly, accurately, and easily distribute disease information to the public.
Through our analysis,
This study has some limitations. First, despite using big data to analyze the phenomenon of public reaction toward the COVID-19 pandemic, some noise or spam remains in the data set; the SocialHeat tool could not completely filter these out due to technology limitations and the complexities of natural language. Though natural language has been applied and innovated daily in SocialHeat’s tool, some texts or paragraphs containing incorrect grammar, teen code, dialects, etc, could not be processed or categorized. It is also important to note that although the data set was pulled from diverse sources like Facebook, YouTube, news sites, etc, the observed format was text only. This means that other formats such as video with text or audio captions or images with textboxes were not analyzed by the SocialHeat tool. Hence, this led to a shortage in the final data set results such as sentiments categorized, extracted top sources, and extracted top posts.
Additionally, due to privacy policies, the data set can only collect data that is installed in public mode, especially for data obtained from social networking platforms like YouTube, Instagram, and Facebook. Moreover, because the data collection time is quite long (11 months), the amount of data poured into the system is large and requires a substantial amount of time for the system to process noise and spam, and give statistical results. This led to a situation in which we wanted to analyze the
Finally, we have almost 38 million data in total, which the system could not process all at once due to technical limitations. Therefore, we could not extract top posts by mentions, top sources by mentions, or overall sentiment of all sources.
The topics discussed on the COVID-19 issue are varied. The classification of content groups as we propose in the study is still limited when it is impossible to analyze the public’s emotional index for each type of topic. Understanding the feelings of the community on specific topics related to the COVID-19 topic can help the government and stakeholders come up with precise and meticulous guidance on disease reactions. Therefore, we suggest that researchers focus on analyzing the public’s sentiment index for each type of topic that the public is discussing to come up with appropriate ideas and options to support the medical information management in pandemic times.
Through our research, we found that using different types of information sources can be effective in different pandemic phases. The same goes for pandemic-related content types. We also highlighted hot spots of public concern regarding the COVID-19 pandemic. These results can help governments or health educators communicate pandemic prevention guidelines more effectively to the public. This is significant not only for prevention during the current COVID-19 pandemic but also could serve as a useful reference for the health crisis management field for potential diseases in the future.
Applying big data in infodemiology studies opens opportunities for getting better insights into a public reaction toward pandemics and related events. The government should take advantage of social platforms to effectively communicate health information, quickly address fake news, and give real-time response to the hot issues that the public needs to know during the pandemic. To achieve those goals, we suggest three key points to help government and stakeholders have better communication with the public during crisis events like the COVID-19 pandemic:
Applying artificial intelligence tools in analyzing big data from social media platforms to collect public insights, determine appropriate cooperation channels in spreading news and guidelines, and effectively communicate about health information and instructions
Promoting an official account of the Ministry of Health on different social media platforms to form the public’s habit of updating news from official sources, avoiding
Collaborating with popular community and KOLs’ fan pages to spread information faster and wider to various reader segments
Big data is also meaningful for infodemiology studies. Applying big data allows researchers to have a wider view and easily compare the results across countries, regions, races, or cultures and lead to more research ideas such as descriptive studies or predicting public sentiments or public reactions about the pandemic.
Total mention trend line from December 1, 2019, to November 13, 2020.
application programming interface
key opinion leader
severe acute respiratory syndrome
World Health Organization
This project received support from YouNet Group in big data crawling and analysis.
None declared.