Corpus card · Sessions 2–4 and 6
Presidential speeches, 100- and 300-speech subsets
South Korean presidential speeches by Lee Myung-bak, Park Geun-hye, and Moon Jae-in, prepared for the weekly demos. Sessions 2 and 3 use 100 speeches. Session 4 uses 300.
presidential-speeches-100.csv, 100 speeches (0.5 MB, UTF-8)
presidential-speeches-300.csv, 300 speeches (1.5 MB, UTF-8)
Codebook, the same columns in both files
Save the files in
data/ under these names in your course repository. You may commit them. Use this card to fill in data/README.md.
Source and selection
The speeches come from the speech records of the Presidential Archive of the Republic of Korea (대통령기록관), through the instructor’s NLP Corpora for Korean Studies catalog. One row is one speech. division_number is the archive’s record number and is unique in the file.
| President | Speeches | Dates | Undated |
|---|---|---|---|
| 이명박 (Lee Myung-bak) | 37 | 2008.03.19 – 2012.06.22 | 7 |
| 박근혜 (Park Geun-hye) | 20 | 2013.05.21 – 2016.10.26 | 0 |
| 문재인 (Moon Jae-in) | 43 | 2017.07.11 – 2022.03.17 | 0 |
| Total | 100 | 7 |
The file is a random selection from a larger sample of democratic-era speeches (seed 42). The procedure for the final 100 was not recorded. No preprocessing or correction was applied. The text is Korean only.
The 300-speech file
The 300-speech file keeps all 100 speeches above and adds 200 more, so that each president has 100.
| President | Speeches | Dates | Undated |
|---|---|---|---|
| 이명박 (Lee Myung-bak) | 100 | 2008.03.19 – 2012.08.15 | 16 |
| 박근혜 (Park Geun-hye) | 100 | 2013.03.27 – 2016.10.26 | 0 |
| 문재인 (Moon Jae-in) | 100 | 2017.06.06 – 2022.03.17 | 0 |
| Total | 300 | 16 |
The 200 added speeches are a random draw for each president (seed 2026) from the same archive file. Only speeches of 400 to 10,000 characters were eligible, the range of the 100. The archive holds many of Lee’s speeches twice, under two record numbers, so a drawn speech was skipped when it repeated a speech already in the file. Rows are ordered by president, then date.
The mix of speech types still differs by president. Most of Lee’s (64) and Park’s (79) speeches are 기념사. Moon’s are spread over 회의, 기타, 성명/담화문, and 축사.
Columns
| Column | Contents |
|---|---|
division_number |
Archive record number |
president |
President, in Korean |
title |
Speech title, in Korean |
date |
Date as YYYY.MM.DD, empty for 7 speeches |
location |
국내 (domestic), 국외 or 해외 (both overseas) |
kind |
Speech type, such as 기념사 (commemorative) or 회의 (meeting remarks). The 300-speech file adds 신년사 (New Year address) |
speech_text |
Full speech text |
Open the file with File, then connect Corpus. The Corpus widget selects title as the text feature by default. Set it to speech_text.
Known limitations
- A difference between presidents may reflect speech type, audience, date range, document length, archival coverage, or an unbalanced selection.
- The mix of speech types differs by president. Most of Lee’s and Park’s speeches are 기념사. Most of Moon’s are 회의 and 기타.
- In the 100-speech file, group sizes range from 20 to 43 speeches. The 300-speech file has 100 per president. Neither represents all presidential language.
- Speeches missing from the archive are missing here.
Permitted use
The texts are public presidential records. Use the file for this course. The file is small, so you may commit it to your private course repository in data/. Course submissions may include counts, figures, tables, and short quotations with the speech title and date. Do not publish the full texts or the processed text column.
Cite the corpus as follows. Presidential speeches, 대통령기록관 (Presidential Archive of the Republic of Korea), prepared subset presidential-speeches-100 v1, BA3 Text as Data 2026–2027. For Session 4, cite presidential-speeches-300 v1 in the same way.