I extracted the keywords from my datasets using the online tool Tagxedo(http://www.
com/), which I highly recommend other data scientists to utilize because of its convenience, and organized them in the form of yellow ribbons(the symbol of the tragedy of Sewol).
AnalysisAfter examining the changes in the word clouds throughout the years, I was able to identify three significant traits related to the Sewol incident.
Examining HistoryListing the relevant events inferred from the word cloudsFirst of all, these word clouds could be used as valuable resources for determining the relevant events that occurred over the years.
For example from words like ‘세월호 특별법’ (Sewol special law) and ‘유가족’ (surviving families) in the 2014 word cloud, we could infer that the public’s sympathy for the deceased children’s surviving family members was strong and the idea of the Sewol special law(a law for investigating the cause of the incident, and compensating for the remaining families’ loss) was introduced during that time.
In 2016, words like ‘특조위’(special investigation committee) and ‘박근혜’ (then-president Park) indicates that the Sewol investigation committee was activated and the public was angry at then-president Park for her disinterest toward the Sewol incident.
I was also able to know that the Sewol disaster was closely linked to the 2017 Korean presidential election as the names of the presidential candidates frequently appeared in the word cloud.
Political Exploitation of the Tragedy of SewolEvidence that the tragedy of Sewol is being politically exploitedSecondly, while I was examining the word clouds, I discovered something peculiar.
While the number and size of words that display sympathy such as ‘유가족’(surviving families), ‘함께’(together), and ‘희생자’(victims) were decreasing over time, the number and size of words related to politics such as the names of politicians or political parties kept increasing.
It was especially notable that the names of politicians previously non-existent had begun to appear in large numbers during the 2017 presidential election.
So what could this phenomenon tell us?.I believe that this indicates that the tragedy of Sewol is being politically exploited by political figures and parties that strive to further their agenda.
Perhaps this could be the reason why some people are expressing hostility toward the Sewol incident; because they are tired of seeing politicians exploiting the Sewol tragedy to gain attention.
Decrease in the public’s sympathy toward the Sewol incidentFinally, I visualized the change in the number of tweets related to Sewol each year as a bar graph in order to examine whether the public’s attention toward the incident is consistent over time.
As you can see from the graph above, it turns out that the public’s interest in Sewol has been dwindling over time.
Approximately 10000 tweets were uploaded during 2014, but the number decreased abruptly in the following year and has been gradually decreasing ever since.
The fact that there were only about hundred tweets around recent years made me a bit afraid of the indifference of society toward the deaths of over a hundred students.
My ThoughtsIn conclusion, it is evident that the public’s sincere sympathy for the Sewol incident is fading due to various factors including political exploitation of the event.
As you can see, the results of my data analyses are pretty bleak.
So, what are we going to do about it?.Well unfortunately, I don’t have a definite solution for this conundrum.
My focus is more concentrated on the novel method of analyzing public perception that I’ve discovered while conducting this project.
By analyzing the social media data related to Sewol, I learned that data science could not only be used for computer-y tasks such as optimizing numerical values, but also for gaining more insight on public issues.
I believe that what my project has accomplished is that, it successfully portrayed the social changes driven by the tragedy of Sewol.
Perhaps by documenting these changes, we are able to hold on to the memory of the deceased a little longer.
Thanks for reading!.Here’s my Github page that contains the resources utilized in this project in case you want to try it yourselves!^^ https://github.