Sri Lankan Parliamentary Speeches in Three Languages, 2017–2026

19,553 parliamentary speeches from the Sri Lankan Parliament, spanning 2017 to 2026. The corpus is trilingual, containing speeches in Sinhala, Tamil, and English, including code-mixed utterances. It was constructed by author himath-nimpura for a thesis on multilingual topic modeling.

Use Cases

• Multilingual topic modeling based on the unsupervised topic cluster labels generated by the BiTopic pipeline
• Analyzing political discourse and speaker contributions based on the associated speaker names and dates
• Studying code-switching and language use in formal settings based on the trilingual and code-mixed content
• Training or evaluating NLP models for Sinhala and Tamil languages based on the parliamentary speech corpus

Strengths

• Contains 19,553 individual speech records, providing substantial textual data
• Covers a nine-year time range from 2017 to 2026, offering longitudinal analysis potential
• Includes trilingual content (Sinhala, Tamil, English) and code-mixed utterances, which is relatively rare
• Each speech is associated with metadata like speaker name, date, and an unsupervised topic cluster label

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is known, but other structural details like file formats and specific column names are unknown
• The dataset's license is unspecified, which may restrict certain uses

Provenance

Source

Sri Lankan Parliament (Hansard)

Collection Method

Constructed and used in the thesis 'Multilingual Topic Modeling of Sri Lankan...' by himath-nimpura

Time Range

2017–2026

Freshness

Last updated 2026-04-10 06:55:36

Geography

Sri Lanka