AfriSpeech/afri-topics

AfriTopics

A topic-classification layer over AfriSpeech/africa-corpus (Bible text) and AfriSpeech/african-corpus-jw (JW.org publications), classifying the English anchor text into a fixed topic taxonomy with Gemini, then propagating the label to every aligned African-language translation via the shared verse_key.

Topics

Every example carries exactly one of these 10 topic labels (topic column):

Topic	Covers
agriculture	farming, crops, livestock, fisheries, food supply, food prices and scarcity
health	illness, disease, hospitals, medicine, public health, mental health
governance	government, politics, elections, law, public administration, corruption, leadership
economy	finance, trade, business, employment, income, poverty, cost of living
education	schools, students, teachers, universities, literacy, curriculum
technology	software, computers, telecoms, internet, AI, engineering and science advances
infrastructure	roads, power, water and sanitation, transport, housing, construction
environment	climate, natural disasters, pollution, conservation, sustainable resource use
culture	communities, religion and belief, languages, arts, traditions, family, identity
security	crime, conflict, violence, policing, peacebuilding, community safety

How labels were assigned

Every item is categorized into one of the 10 topics above — there is no none/catch-all option. An earlier pass allowed a none label for content with no clear real-world subject matter (narrative, genealogy, poetry, fragments, captions); on review that made the dataset less useful, so those items were re-classified with none removed from the label set, forcing a best-fit real-topic decision even where the fit is loose (e.g. Genesis 1's creation narrative → environment). Any remaining un-reclassified items were filtered out, ensuring that 100% of items in the published dataset carry a concrete topic label.

Structure

• english/english_bible.csv — verse_key, eng, topic (28711 verses)
• bible/{Name}_{iso3}_v{N}.csv — verse_key, local, topic, version; every translation version of each language is kept (693 languages, 949 files)
• english/english_jw.csv — verse_key, eng, topic (386427 paragraphs)
• jw/{Name}_{code}.csv — verse_key, local, topic, version, one file per JW-corpus African language (135 languages)

The Bible and JW portions use different key namespaces (BOOK.chapter.verse vs docid#pid) and are kept as separate files rather than merged.

Loading

Each language is a config named {Name}_{iso3} (e.g. Twi_twi), with bible and/or jw splits (load_dataset("AfriSpeech/afri-topics", "Twi_twi", split="bible")). The english config holds the English source text and topic for each split.

Method

1. Pulled the English anchor text (english_cache.csv) from each source dataset.
2. For the Bible corpus, classified verses with Gemini (gemini-2.5-flash).
3. For the JW corpus, first restricted to paragraphs that have at least one African-language parallel translation anywhere in the corpus (paragraphs with zero parallels can't be used for a parallel dataset anyway), then classified those with Gemini (gemini-2.5-flash-lite, thinking disabled).
4. Second pass: items labeled none in step 2/3 were re-classified with none removed from the label set, forcing the model to pick whichever of the 10 real topics fits best. Items not re-classified were omitted so all retained examples carry a valid topical label.
5. Joined the resulting verse_key -> topic labels onto each language's per-language file, covering every language in the Bible corpus.
6. Bible-corpus languages with several translation versions keep all of them; the version column (source version id, e.g. v1631) tells rows apart within a language's bible split.

Caveats

• Labels are LLM judgments on the English text only, not human-validated, and not verified against the African-language text itself — treat them as weak/silver labels. A meaningful fraction were forced best-fit choices on content with no strong real-world topical grounding (narrative, genealogy, fragments) — see the taxonomy note above.
• An earlier attempt classified this data with a pure sentence-embedding nearest-neighbor approach (BAAI/bge-large-en-v1.5 against category description embeddings); it was abandoned after inspection showed low, non-discriminating confidence scores and frequent misclassification of fragments/captions. Gemini was used instead for both corpora.
• Category distribution is uneven and reflects the source material: Bible narrative skews toward culture and security (religious/war content); JW literature covers the other topics more evenly but still skews toward culture.

Source & license

Bible text is derived from public Bible translations via YouVersion (see AfriSpeech/africa-corpus for terms). JW text is an alignment index over publicly accessible jw.org pages, content (c) Watch Tower Bible and Tract Society of Pennsylvania (see AfriSpeech/african-corpus-jw for terms). This repository's own contribution (the topic labels and the joined structure) is released under the same terms as the two source datasets.