CongoNames Corpus: A Large-Scale Labeled Dataset of Congolese Personal Names
Fuente:
Zenodo
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Recurso digital |
| Lingua: | francese |
| Pubblicazione: |
Zenodo
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866901577139224576 |
|---|---|
| author | Tshabu Ngandu, Bernard Amaury, Cansa Merlec, Mpyana |
| author_facet | Tshabu Ngandu, Bernard Amaury, Cansa Merlec, Mpyana |
| contents | <p><span>Personal names carry cultural and linguistic identity, yet most African countries lack large-</span><span>scale, structured name datasets suitable for natural language processing (NLP) research</span> <span>and computational social science. We present C</span><span>ONGO</span><span>N</span><span>AMES</span><span>, the first large-scale corpus of</span> <span>personal names from the Democratic Republic of the Congo (DRC), derived from publicly</span> r<span>eleased national secondary-school examination</span> <span>palomàres</span> <span>(result lists) published annually</span> <span>by the DRC Ministry of Education. The corpus comprises</span> <span>8,053,983</span> <span>name records spanning</span> <span>16 examination years (2008–2023) across 12 provinces and 304 sub-provincial regions, each</span> <span>enriched with a reported sex marker (</span><span>M</span><span>/</span><span>F</span><span>) and regional provenance metadata. We describe</span> <span>a fully deterministic, layered processing pipeline (bronze–silver–gold architecture) that</span> <span>converts raw PDF documents into structured CSV datasets without manual annotation</span> <span>or machine-learning-based inference. The dataset is validated against school-level census</span> <span>counts extracted from the same source PDFs, yielding extraction error rates below 2% for all</span> <span>years except 2023 (7.81%, flagged due to a layout change). Descriptive analyses document</span> <span>name length and token-count distributions, character-level</span> <span>n</span><span>-gram profiles, provincial </span><span>diversity indices, and inter-provincial name-inventory overlap, collectively establishing</span> <span>the dual linguistic origin—locally rooted Bantu components and Christian/French-origin</span> <span>components—that characterizes modern Congolese naming practice. The dataset, pro</span><span>cessing code, and documentation are released openly to support research in African NLP,</span> <span>onomastics, and computational social science.</span></p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_19809985 |
| institution | Zenodo |
| language | fra |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | CongoNames Corpus: A Large-Scale Labeled Dataset of Congolese Personal Names Tshabu Ngandu, Bernard Amaury, Cansa Merlec, Mpyana African Languages Democratic Republic of Congo Congo Names Natural Language Processing Onomastics <p><span>Personal names carry cultural and linguistic identity, yet most African countries lack large-</span><span>scale, structured name datasets suitable for natural language processing (NLP) research</span> <span>and computational social science. We present C</span><span>ONGO</span><span>N</span><span>AMES</span><span>, the first large-scale corpus of</span> <span>personal names from the Democratic Republic of the Congo (DRC), derived from publicly</span> r<span>eleased national secondary-school examination</span> <span>palomàres</span> <span>(result lists) published annually</span> <span>by the DRC Ministry of Education. The corpus comprises</span> <span>8,053,983</span> <span>name records spanning</span> <span>16 examination years (2008–2023) across 12 provinces and 304 sub-provincial regions, each</span> <span>enriched with a reported sex marker (</span><span>M</span><span>/</span><span>F</span><span>) and regional provenance metadata. We describe</span> <span>a fully deterministic, layered processing pipeline (bronze–silver–gold architecture) that</span> <span>converts raw PDF documents into structured CSV datasets without manual annotation</span> <span>or machine-learning-based inference. The dataset is validated against school-level census</span> <span>counts extracted from the same source PDFs, yielding extraction error rates below 2% for all</span> <span>years except 2023 (7.81%, flagged due to a layout change). Descriptive analyses document</span> <span>name length and token-count distributions, character-level</span> <span>n</span><span>-gram profiles, provincial </span><span>diversity indices, and inter-provincial name-inventory overlap, collectively establishing</span> <span>the dual linguistic origin—locally rooted Bantu components and Christian/French-origin</span> <span>components—that characterizes modern Congolese naming practice. The dataset, pro</span><span>cessing code, and documentation are released openly to support research in African NLP,</span> <span>onomastics, and computational social science.</span></p> |
| title | CongoNames Corpus: A Large-Scale Labeled Dataset of Congolese Personal Names |
| topic | African Languages Democratic Republic of Congo Congo Names Natural Language Processing Onomastics |
| url | https://doi.org/10.5281/zenodo.19809985 |