@inproceedings{czerski-ogrodniczuk-2026-localgovpl,
title = "{L}ocal{G}ov{PL}: A Corpus of Speaker-Attributed {P}olish Local Government Transcripts",
author = "Czerski, Dariusz and
Ogrodniczuk, Maciej",
editor = "Piperidis, Stelios and
Bel, N{\'u}ria and
van den Heuvel, Henk and
Ide, Nancy and
Krek, Simon and
Toral, Antonio",
booktitle = "Proceedings of the Fifteenth Language Resources and Evaluation Conference",
month = may,
year = "2026",
address = "Palma de Mallorca, Spain",
publisher = "ELRA Language Resource Association",
url = "https://aclanthology.org/2026.lrec-1.626/",
doi = "10.63317/23722jebsz2d",
pages = "7883--7893",
abstract = "We present LocalGovPL, a large-scale, speaker-annotated corpus of Polish local government meeting transcripts processed using an automatic two-stage LLM pipeline. The corpus consists of 31,900 sessions from 749 councils recorded between 2018{--}2025 (approximately 391M words). It is released in TEI P5 format with explicit links between utterances and registered participants. We collect transcripts from official local government portals using a dedicated crawler, normalize the text, and apply: (1) LLM-assisted extraction of person names and administrative roles; and (2) attribution of utterances to identified speakers using discourse cues. To evaluate attribution quality, we manually annotate 30 sessions and evaluate five LLM configurations using three evaluation protocols with speaker-aware word error rate (sWER). The strongest system, Gemini-2.5-pro, achieves 3.9{\%} sWER for abstract speaker identification, 4.6{\%} for known participants, and 5.9{\%} for end-to-end processing with relaxed name matching. LocalGovPL enables large-scale analysis of local deliberative discourse and supports research on dialogue modeling, summarization, and political text analysis."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="czerski-ogrodniczuk-2026-localgovpl">
<titleInfo>
<title>LocalGovPL: A Corpus of Speaker-Attributed Polish Local Government Transcripts</title>
</titleInfo>
<name type="personal">
<namePart type="given">Dariusz</namePart>
<namePart type="family">Czerski</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Maciej</namePart>
<namePart type="family">Ogrodniczuk</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Fifteenth Language Resources and Evaluation Conference</title>
</titleInfo>
<name type="personal">
<namePart type="given">Stelios</namePart>
<namePart type="family">Piperidis</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Núria</namePart>
<namePart type="family">Bel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Henk</namePart>
<namePart type="family">van den Heuvel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nancy</namePart>
<namePart type="family">Ide</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="family">Krek</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Antonio</namePart>
<namePart type="family">Toral</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resource Association</publisher>
<place>
<placeTerm type="text">Palma de Mallorca, Spain</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>We present LocalGovPL, a large-scale, speaker-annotated corpus of Polish local government meeting transcripts processed using an automatic two-stage LLM pipeline. The corpus consists of 31,900 sessions from 749 councils recorded between 2018–2025 (approximately 391M words). It is released in TEI P5 format with explicit links between utterances and registered participants. We collect transcripts from official local government portals using a dedicated crawler, normalize the text, and apply: (1) LLM-assisted extraction of person names and administrative roles; and (2) attribution of utterances to identified speakers using discourse cues. To evaluate attribution quality, we manually annotate 30 sessions and evaluate five LLM configurations using three evaluation protocols with speaker-aware word error rate (sWER). The strongest system, Gemini-2.5-pro, achieves 3.9% sWER for abstract speaker identification, 4.6% for known participants, and 5.9% for end-to-end processing with relaxed name matching. LocalGovPL enables large-scale analysis of local deliberative discourse and supports research on dialogue modeling, summarization, and political text analysis.</abstract>
<identifier type="citekey">czerski-ogrodniczuk-2026-localgovpl</identifier>
<identifier type="doi">10.63317/23722jebsz2d</identifier>
<location>
<url>https://aclanthology.org/2026.lrec-1.626/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>7883</start>
<end>7893</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T LocalGovPL: A Corpus of Speaker-Attributed Polish Local Government Transcripts
%A Czerski, Dariusz
%A Ogrodniczuk, Maciej
%Y Piperidis, Stelios
%Y Bel, Núria
%Y van den Heuvel, Henk
%Y Ide, Nancy
%Y Krek, Simon
%Y Toral, Antonio
%S Proceedings of the Fifteenth Language Resources and Evaluation Conference
%D 2026
%8 May
%I ELRA Language Resource Association
%C Palma de Mallorca, Spain
%F czerski-ogrodniczuk-2026-localgovpl
%X We present LocalGovPL, a large-scale, speaker-annotated corpus of Polish local government meeting transcripts processed using an automatic two-stage LLM pipeline. The corpus consists of 31,900 sessions from 749 councils recorded between 2018–2025 (approximately 391M words). It is released in TEI P5 format with explicit links between utterances and registered participants. We collect transcripts from official local government portals using a dedicated crawler, normalize the text, and apply: (1) LLM-assisted extraction of person names and administrative roles; and (2) attribution of utterances to identified speakers using discourse cues. To evaluate attribution quality, we manually annotate 30 sessions and evaluate five LLM configurations using three evaluation protocols with speaker-aware word error rate (sWER). The strongest system, Gemini-2.5-pro, achieves 3.9% sWER for abstract speaker identification, 4.6% for known participants, and 5.9% for end-to-end processing with relaxed name matching. LocalGovPL enables large-scale analysis of local deliberative discourse and supports research on dialogue modeling, summarization, and political text analysis.
%R 10.63317/23722jebsz2d
%U https://aclanthology.org/2026.lrec-1.626/
%U https://doi.org/10.63317/23722jebsz2d
%P 7883-7893
Markdown (Informal)
[LocalGovPL: A Corpus of Speaker-Attributed Polish Local Government Transcripts](https://aclanthology.org/2026.lrec-1.626/) (Czerski & Ogrodniczuk, LREC 2026)
ACL