Corpus card · Session 5
Presidential tweets, Moon Jae-in
Tweets from Moon Jae-in's official account, 2012 to 2020, prepared for classification and sentiment analysis.
presidential-tweets-moon.csv (1.2 MB, UTF-8) · codebook
Save the file as
data/private/presidential-tweets/presidential-tweets-moon.csv in your course repository. Do not commit it.
Source and selection
The tweets come from Moon Jae-in’s official Twitter account (moonriver365), from a public dataset on Kaggle, through the instructor’s NLP Corpora for Korean Studies catalog. One row is one tweet. The 20 tweets with no text in the source are removed.
| Period | Dates | Tweets |
|---|---|---|
| Before the presidency | 2012-01-01 – 2016-11-30 | 1,972 |
| Transition | 2016-12-01 – 2017-05-09 | 392 |
| Presidency | 2017-05-10 – 2020-06-15 | 764 |
| Total | 3,128 |
Almost half of all tweets date from 2012, when Moon was a presidential candidate. He tweeted far less once in office. The text is mostly Korean. About 120 tweets are mainly English or Spanish, usually diplomatic messages.
Columns
| Column | Contents |
|---|---|
username |
Account handle, always moonriver365 |
tweet_date, tweet_time |
Date and time of the tweet |
text |
Full tweet text, including links, hashtags, and emoji |
favorites, retweets |
Counts at the time of collection |
link |
URL of the original tweet |
tweet_year |
Year, 2012 to 2020 |
period3 |
pre_presidency, transition, or presidency |
Open the file with File, then connect Corpus. If the Corpus widget picks a different text feature, set it to text.
Known limitations
- Engagement counts do not measure public opinion.
- An official account is only part of a president’s public communication.
- Deleted, media-only, and missing posts are not in the file.
- Tweets are short and use platform conventions, so they are not comparable to speeches without care.
Permitted use
Use the file for this course. Keep it out of your repository, which ignores data/private/. Course submissions may include counts, figures, tables, and short quotations with the tweet date. Do not publish the full texts.
Cite the corpus as follows. Moon Jae-in tweets (moonriver365), 2012–2020, prepared subset presidential-tweets-moon v1, BA3 Text as Data 2026–2027.