Corpus card · Session 5
National Assembly sentences on North Korea and Japan
Sentences from the minutes of the National Assembly, 2000 to 2026, that name North Korea or Japan, prepared for sentiment analysis and clustering.
assembly-nk-japan.csv (1.4 MB, UTF-8) · codebook
Save the file in the
data/ folder of your course repository.
Source and selection
The sentences come from the minutes of every committee and plenary sitting of the National Assembly, 2000 to 2026, published at record.assembly.go.kr. One row is one sentence that names 북한 (North Korea) or 일본 (Japan). Sentences that name both are left out, and so are sentences shorter than 30 or longer than 300 characters.
The file holds 65 sentences per keyword per year, drawn at random, so that no year dominates.
| Keyword | Sentences |
|---|---|
| 북한 | 1,495 |
| 일본 | 1,495 |
| Total | 2,990 |
Legislators speak 2,362 of the sentences and ministers 628. Of the legislators’ sentences, 933 come from the ruling party, 1,359 from the opposition, and 70 from independents.
Columns
| Column | Contents |
|---|---|
source_id |
Minutes document and turn the sentence comes from |
date, year |
Sitting date and year |
president, president_camp |
President in office, and progressive or conservative |
assembly_term |
National Assembly term (대), 16 to 22 |
committee_group |
Committee policy area, harmonised across renamings |
speaker, position, role |
Speaker name and position as printed, and legislator or government |
party, bloc, ruling_opposition |
Party on the sitting date, its party family, and ruling party, opposition, or independent |
keyword |
북한 or 일본 |
sentence |
The sentence as the minutes print it, not preprocessed |
nouns |
The sentence after the course Kiwi script with its default settings, for clustering |
Open the file with File, then connect Corpus. Use sentence as the text feature for sentiment analysis and nouns for clustering.
Known limitations
- The minutes from 2008 to mid-2012 exist only as scans and are not in the file.
- A sentence is cut from a longer speech turn, so it can lose the context that gives it its tone.
- Members speak for an audience, in committee or in plenary. The minutes record what was said, not what members think.
- Ministers answer questions, so their sentences differ in kind from the legislators’ questions.
Permitted use
Use the file for this course. Course submissions may include counts, figures, tables, and short quotations with the sitting date and speaker position.
Cite the corpus as follows. National Assembly minutes (record.assembly.go.kr), 2000–2026, prepared subset assembly-nk-japan v1, BA3 Text as Data 2026–2027.