EPDiscourse: European Parliament Plenary Discourse Corpus, 1999–2026

EPDiscourse is a harmonised longitudinal corpus of 604,927 European Parliament plenary speech and written-contribution records covering European Parliament terms 5 through 10 and the period from 20 July 1999 to 18 June 2026. It combines 574,119 records inherited from the European Parliament component of ParlLawSpeech with 30,808 records obtained from official European Parliament data for the beginning of term 10. The public dataset contains 21 variables, including stable identifiers, dates, speaker and role indicators, harmonised European Parliament political-group information, agenda information, multilingual original text, unified analytical English text, official English text where available, translation provenance, and source URLs. Analytical English text is available for 604,905 records; 22 records lack text in all three public text fields. The corpus is intended for computational text analysis, descriptive research, and statistical analysis of European parliamentary discourse. Researchers should account for source heterogeneity and the non-authoritative character of machine-generated English translations.