Dataset Viewer
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
word: string
position: int64
language: string
languageName: string
script: string
integrationClass: string
meaning_en: string
etymology: string
notes: string
entryId: int64
comments: string
customFields: string
id: int64
root: string
rootLanguage: string
rootLanguageFamily: string
prefix: string
suffix: string
pos: string
lemma: string
meaning_contextual: string
meaningShift: string
semanticField: string
connotation: string
register: string
domain: string
yt_inputMode: string
ns_speakerLangs: string
filename: string
geo: struct<country: string, city: string, district: string, street: string, postcode: string, raw: strin (... 2 chars omitted)
child 0, country: string
child 1, city: string
child 2, district: string
child 3, street: string
child 4, postcode: string
child 5, raw: string
yt_videoUrl: string
fileSize: string
yt_speakerNativeLang: string
error_source: string
editHistory: list<item: null>
child 0, item: null
source_description: string
detectedLanguages: list<item: string>
child 0, item: string
ocrConfidence_source: string
yt_speakerName: string
languageNames: list<item: string>
child 0, item: string
error: string
sourceType: string
entryType: string
ns_speakerName: string
scripts: list<item: string>
child 0, item: string
ns_generationContext: string
entryType_source: string
yt_channelUrl: string
switchTrigger_source: string
transcription: string
detectedLanguages_source: string
linguisticNotes: string
scripts_source: string
ns_recordingContext: st
...
xif_source: string
historicalScript: string
type: string
geo_source: string
ns_consentRecorded: string
switchType: string
hasText: bool
transliteration_source: string
yt_licence: string
source_type_label: string
linguisticNotes_source: string
matrixLang: string
transliteration: string
status: string
translation: struct<en: string>
child 0, en: string
yt_speakerRole: string
collectedBy_source: string
ocrConfidence: string
ocrText: string
yt_lessonTopic: string
transcription_source: string
switchType_source: string
historicalScript_source: string
createdAt: string
imageDescription_source: string
zone: string
audienceLanguages: list<item: string>
child 0, item: string
collectedBy: string
notes_source: string
languageNames_source: string
hasCodeSwitching_source: string
ns_speakerNeighbourhood: string
ns_speakerEducation: string
yt_duration: string
audienceLanguages_source: string
yt_permissionType: string
exif: struct<datetime: string, make: string, model: string, lat: double, lon: double, altitude: string, da (... 24 chars omitted)
child 0, datetime: string
child 1, make: string
child 2, model: string
child 3, lat: double
child 4, lon: double
child 5, altitude: string
child 6, datetimeDigitized: string
ns_speakerAge: string
yt_channelName: string
ocrText_source: string
imageDescription: string
domain_source: string
yt_permissionDate: timestamp[s]
translation_source: string
hasCodeSwitching: bool
yt_publishedDate: string
matrixLang_source: string
rawText: string
to
{'filename': Value('string'), 'type': Value('string'), 'exif': {'datetime': Value('string'), 'make': Value('string'), 'model': Value('string'), 'lat': Value('float64'), 'lon': Value('float64'), 'altitude': Value('string'), 'datetimeDigitized': Value('string')}, 'exif_source': Value('string'), 'geo': {'country': Value('string'), 'city': Value('string'), 'district': Value('string'), 'street': Value('string'), 'postcode': Value('string'), 'raw': Value('string')}, 'geo_source': Value('string'), 'zone': Value('string'), 'zone_source': Value('string'), 'collectedBy': Value('string'), 'status': Value('string'), 'createdAt': Value('string'), 'editHistory': List(Value('null')), 'id': Value('int64'), 'ocrText': Value('string'), 'imageDescription': Value('string'), 'detectedLanguages': List(Value('string')), 'languageNames': List(Value('string')), 'scripts': List(Value('string')), 'translation': {'en': Value('string')}, 'transliteration': Value('string'), 'domain': Value('string'), 'entryType': Value('string'), 'matrixLang': Value('string'), 'hasCodeSwitching': Value('bool'), 'switchType': Value('string'), 'switchTrigger': Value('string'), 'linguisticNotes': Value('string'), 'ocrText_source': Value('string'), 'imageDescription_source': Value('string'), 'detectedLanguages_source': Value('string'), 'languageNames_source': Value('string'), 'scripts_source': Value('string'), 'translation_source': Value('string'), 'transliteration_source': Value('string'), 'domain_source': Value('string'), '
...
ode=True), 'transcription': Value('string'), 'transcription_source': Value('string'), 'ocrConfidence': Value('string'), 'ocrConfidence_source': Value('string'), 'hasText': Value('bool'), 'hasText_source': Value('string'), 'audienceLanguages': List(Value('string')), 'historicalScript': Value('string'), 'audienceLanguages_source': Value('string'), 'historicalScript_source': Value('string'), 'sourceType': Value('string'), 'yt_videoTitle': Value('string'), 'yt_videoUrl': Value('string'), 'yt_publishedDate': Value('string'), 'yt_duration': Value('string'), 'yt_speakerName': Value('string'), 'yt_speakerRole': Value('string'), 'yt_speakerNativeLang': Value('string'), 'yt_languageFocus': Value('string'), 'yt_lessonTopic': Value('string'), 'yt_permissionGrantedBy': Value('string'), 'yt_permissionDate': Value('timestamp[s]'), 'yt_permissionType': Value('string'), 'yt_channelName': Value('string'), 'yt_channelUrl': Value('string'), 'yt_licence': Value('string'), 'source_type_label': Value('string'), 'source_description': Value('string'), 'fileSize': Value('string'), 'ns_speakerName': Value('string'), 'ns_speakerAge': Value('string'), 'ns_speakerNeighbourhood': Value('string'), 'ns_speakerEducation': Value('string'), 'ns_speakerLangs': Value('string'), 'ns_generationContext': Value('string'), 'ns_recordingContext': Value('string'), 'ns_consentRecorded': Value('string'), 'error': Value('string'), 'error_source': Value('string'), 'rawText': Value('string'), 'yt_inputMode': Value('string')}
because column names don't match
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
return get_rows(
dataset=dataset,
...<4 lines>...
column_names=column_names,
)
File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
return func(*args, **kwargs)
File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
File "/src/services/worker/src/worker/utils.py", line 478, in safe_iter
yield from ds.decode(False) if ds.features else ds
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2818, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2355, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2380, in _iter_arrow
for key, pa_table in self.ex_iterable._iter_arrow():
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
for key, pa_table in self.generate_tables_fn(**gen_kwags):
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2369, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
word: string
position: int64
language: string
languageName: string
script: string
integrationClass: string
meaning_en: string
etymology: string
notes: string
entryId: int64
comments: string
customFields: string
id: int64
root: string
rootLanguage: string
rootLanguageFamily: string
prefix: string
suffix: string
pos: string
lemma: string
meaning_contextual: string
meaningShift: string
semanticField: string
connotation: string
register: string
domain: string
yt_inputMode: string
ns_speakerLangs: string
filename: string
geo: struct<country: string, city: string, district: string, street: string, postcode: string, raw: strin (... 2 chars omitted)
child 0, country: string
child 1, city: string
child 2, district: string
child 3, street: string
child 4, postcode: string
child 5, raw: string
yt_videoUrl: string
fileSize: string
yt_speakerNativeLang: string
error_source: string
editHistory: list<item: null>
child 0, item: null
source_description: string
detectedLanguages: list<item: string>
child 0, item: string
ocrConfidence_source: string
yt_speakerName: string
languageNames: list<item: string>
child 0, item: string
error: string
sourceType: string
entryType: string
ns_speakerName: string
scripts: list<item: string>
child 0, item: string
ns_generationContext: string
entryType_source: string
yt_channelUrl: string
switchTrigger_source: string
transcription: string
detectedLanguages_source: string
linguisticNotes: string
scripts_source: string
ns_recordingContext: st
...
xif_source: string
historicalScript: string
type: string
geo_source: string
ns_consentRecorded: string
switchType: string
hasText: bool
transliteration_source: string
yt_licence: string
source_type_label: string
linguisticNotes_source: string
matrixLang: string
transliteration: string
status: string
translation: struct<en: string>
child 0, en: string
yt_speakerRole: string
collectedBy_source: string
ocrConfidence: string
ocrText: string
yt_lessonTopic: string
transcription_source: string
switchType_source: string
historicalScript_source: string
createdAt: string
imageDescription_source: string
zone: string
audienceLanguages: list<item: string>
child 0, item: string
collectedBy: string
notes_source: string
languageNames_source: string
hasCodeSwitching_source: string
ns_speakerNeighbourhood: string
ns_speakerEducation: string
yt_duration: string
audienceLanguages_source: string
yt_permissionType: string
exif: struct<datetime: string, make: string, model: string, lat: double, lon: double, altitude: string, da (... 24 chars omitted)
child 0, datetime: string
child 1, make: string
child 2, model: string
child 3, lat: double
child 4, lon: double
child 5, altitude: string
child 6, datetimeDigitized: string
ns_speakerAge: string
yt_channelName: string
ocrText_source: string
imageDescription: string
domain_source: string
yt_permissionDate: timestamp[s]
translation_source: string
hasCodeSwitching: bool
yt_publishedDate: string
matrixLang_source: string
rawText: string
to
{'filename': Value('string'), 'type': Value('string'), 'exif': {'datetime': Value('string'), 'make': Value('string'), 'model': Value('string'), 'lat': Value('float64'), 'lon': Value('float64'), 'altitude': Value('string'), 'datetimeDigitized': Value('string')}, 'exif_source': Value('string'), 'geo': {'country': Value('string'), 'city': Value('string'), 'district': Value('string'), 'street': Value('string'), 'postcode': Value('string'), 'raw': Value('string')}, 'geo_source': Value('string'), 'zone': Value('string'), 'zone_source': Value('string'), 'collectedBy': Value('string'), 'status': Value('string'), 'createdAt': Value('string'), 'editHistory': List(Value('null')), 'id': Value('int64'), 'ocrText': Value('string'), 'imageDescription': Value('string'), 'detectedLanguages': List(Value('string')), 'languageNames': List(Value('string')), 'scripts': List(Value('string')), 'translation': {'en': Value('string')}, 'transliteration': Value('string'), 'domain': Value('string'), 'entryType': Value('string'), 'matrixLang': Value('string'), 'hasCodeSwitching': Value('bool'), 'switchType': Value('string'), 'switchTrigger': Value('string'), 'linguisticNotes': Value('string'), 'ocrText_source': Value('string'), 'imageDescription_source': Value('string'), 'detectedLanguages_source': Value('string'), 'languageNames_source': Value('string'), 'scripts_source': Value('string'), 'translation_source': Value('string'), 'transliteration_source': Value('string'), 'domain_source': Value('string'), '
...
ode=True), 'transcription': Value('string'), 'transcription_source': Value('string'), 'ocrConfidence': Value('string'), 'ocrConfidence_source': Value('string'), 'hasText': Value('bool'), 'hasText_source': Value('string'), 'audienceLanguages': List(Value('string')), 'historicalScript': Value('string'), 'audienceLanguages_source': Value('string'), 'historicalScript_source': Value('string'), 'sourceType': Value('string'), 'yt_videoTitle': Value('string'), 'yt_videoUrl': Value('string'), 'yt_publishedDate': Value('string'), 'yt_duration': Value('string'), 'yt_speakerName': Value('string'), 'yt_speakerRole': Value('string'), 'yt_speakerNativeLang': Value('string'), 'yt_languageFocus': Value('string'), 'yt_lessonTopic': Value('string'), 'yt_permissionGrantedBy': Value('string'), 'yt_permissionDate': Value('timestamp[s]'), 'yt_permissionType': Value('string'), 'yt_channelName': Value('string'), 'yt_channelUrl': Value('string'), 'yt_licence': Value('string'), 'source_type_label': Value('string'), 'source_description': Value('string'), 'fileSize': Value('string'), 'ns_speakerName': Value('string'), 'ns_speakerAge': Value('string'), 'ns_speakerNeighbourhood': Value('string'), 'ns_speakerEducation': Value('string'), 'ns_speakerLangs': Value('string'), 'ns_generationContext': Value('string'), 'ns_recordingContext': Value('string'), 'ns_consentRecorded': Value('string'), 'error': Value('string'), 'error_source': Value('string'), 'rawText': Value('string'), 'yt_inputMode': Value('string')}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Code-Switching in Post-Soviet Urban Space: A Linguistic Landscape Dataset from Baku and Northern Azerbaijan
Collector: Sudarshan Manikantan Email: manikantan.sudarshan@gmail.com Collection location: Baku, Guba, Gabala, Khinaliq, Shamakhi — Azerbaijan Collection period: Summer 2026 Licence: CC BY 4.0 Platform: https://sudhirulz.github.io/baku-dataset GitHub: https://github.com/sudhirulz/baku-dataset
Dataset description
This dataset documents multilingual signage and speech in Baku, Azerbaijan and surrounding regions, with a focus on code-switching and language contact phenomena in post-Soviet urban and rural Azerbaijan.
The dataset captures how Azerbaijani, Russian, English, and minority languages (Lezgi, Talysh, Khinaliq, Udi, Tsakhur, Avar, Georgian, Armenian) mix in everyday commercial signage, public announcements, menus, street signs, and spontaneous spoken interaction.
Every entry includes full provenance metadata — GPS coordinates, timestamps, and camera model extracted from EXIF data, reverse-geocoded location via OpenStreetMap, and clear source attribution for all three data sources. Every AI-generated field is labelled with its source so researchers can distinguish machine-generated annotations from researcher-verified data.
Research context
The linguistic landscape of post-Soviet Azerbaijan reflects multiple overlapping historical forces: the Soviet-era dominance of Russian, the post-1991 restoration of Azerbaijani as the state language (accompanied by a shift from Cyrillic to Latin script), and the post-2000 penetration of English through globalisation and digital culture.
Baku's urban signage and speech sit at the intersection of these forces, producing a rich environment of code-switching, morphological hybridisation, and script mixing that is largely undocumented in computational linguistics datasets.
Of particular significance is the inclusion of data from Khinaliq village — home to the Khinaliq language, a language isolate spoken by fewer than 2,000 people with no standard orthography. Documentation of Khinaliq linguistic presence in public space is exceptionally rare in any published dataset.
Dataset statistics
- Total entries: 99
- Processed entries: 94
- Words analysed: 2,500+
- With GPS coordinates: majority of photo entries
- With code-switching: ~43% of entries
- Languages documented: Azerbaijani, Russian, English, Lezgi, Talysh, Khinaliq, Udi, Tsakhur, Avar, Georgian, Armenian, Persian, Turkish
- Collection date: Summer 2026
- Collector: Sudarshan Manikantan
Data sources
1. Linguistic landscape photos (primary)
Photos taken by researcher in public spaces across Azerbaijan — Baku city centre, Baku Old City (İçərişəhər), Guba, Gabala, Khinaliq village, Shamakhi, Gobustan, and en route locations. EXIF metadata preserved — every photo includes GPS coordinates, capture timestamp, and camera model. No personally identifiable information included.
2. Native speaker recordings
Spontaneous and elicited speech from a native Baku Azerbaijani speaker. Recorded with informed verbal consent (consent recorded on audio). Speaker metadata: Baku resident, post-independence generation, multilingual in Azerbaijani and Russian. Data anonymised to initials and age group. Audio transcribed via OpenAI Whisper, linguistically analysed via Claude API.
3. Pedagogical audio — Learn Azerbaijani Today (YouTube)
Selected recordings from Learn Azerbaijani Today (youtube.com/@learnazerbaijanitoday). Used with explicit written permission from Samantha Parker, 7 July 2026. Audio files are not redistributed — only transcriptions and linguistic annotations are included in this dataset.
Dataset files
| File | Description | Rows |
|---|---|---|
entries.jsonl |
One entry per line — photos and audio sources | ~94 |
words.jsonl |
One word per line — deep linguistic analysis | ~2,500 |
entries.csv |
Same as entries.jsonl, CSV format | ~94 |
words.csv |
Same as words.jsonl, CSV format | ~2,500 |
Entry-level fields (entries.jsonl)
| Field | Source | Type | Description |
|---|---|---|---|
id |
System | int | Unique entry ID |
sourceType |
Researcher | string | photo / native_speaker / youtube |
collectedBy |
Researcher | string | Collector name |
filename |
System | string | Original filename |
createdAt |
System | datetime | Entry creation timestamp |
zone |
Researcher + GPS | string | Collection zone |
exif_datetime |
EXIF | datetime | Camera capture timestamp |
exif_lat |
EXIF | float | GPS latitude |
exif_lon |
EXIF | float | GPS longitude |
exif_make |
EXIF | string | Camera manufacturer |
exif_model |
EXIF | string | Camera model |
geo_country |
OpenStreetMap | string | Reverse-geocoded country |
geo_city |
OpenStreetMap | string | Reverse-geocoded city |
geo_district |
OpenStreetMap | string | Reverse-geocoded district |
geo_street |
OpenStreetMap | string | Reverse-geocoded street |
ocrText |
AI (Claude) | string | Verbatim text extracted from image |
transcription |
AI (Claude/Whisper) | string | Audio transcription with language tags |
translation_en |
AI (Claude) | string | English translation |
transliteration |
AI (Claude) | string | Latin-script romanisation |
detectedLanguages |
AI + Researcher | list | ISO language codes |
languageNames |
AI + Researcher | list | Full language names |
scripts |
AI + Researcher | list | Writing systems detected |
domain |
AI + Researcher | string | Functional domain of sign/speech |
entryType |
AI + Researcher | string | Type of sign or speech |
matrixLang |
AI + Researcher | string | Dominant/matrix language code |
hasCodeSwitching |
AI + Researcher | bool | Code-switching present |
switchType |
AI + Researcher | string | Type of code-switching |
switchTrigger |
AI + Researcher | string | Pragmatic trigger for switch |
linguisticNotes |
AI (Claude) | string | Scholarly observations |
notes |
Researcher | string | Researcher field annotations |
comments |
Researcher | dict | Field-level researcher comments |
customFields |
Researcher | dict | Custom researcher annotations |
Word-level fields (words.jsonl)
| Field | Description |
|---|---|
entryId |
Links to parent entry |
word |
Exact word form as it appears in the source |
position |
Position index in source text |
script |
Writing script (Latin, Cyrillic, Arabic, etc.) |
language |
ISO language code |
languageName |
Full language name |
root |
Root or stem of the word |
rootLanguage |
Language the root originates from |
rootLanguageFamily |
Language family of root |
prefix |
Prefix if any, null otherwise |
suffix |
Suffix if any, null otherwise |
integrationClass |
Morphological integration class (see taxonomy below) |
pos |
Part of speech |
lemma |
Base / dictionary form |
meaning_en |
Dictionary meaning in English |
meaning_contextual |
Meaning in this specific sign or text |
meaningShift |
Semantic change from source language, if any |
etymology |
Historical origin of the word |
semanticField |
Semantic domain |
connotation |
Pragmatic connotation |
register |
Sociolinguistic register |
notes |
Linguistically notable observations |
researcherComment |
Free researcher annotation |
Integration class taxonomy
This is the original scholarly contribution of this dataset. Words are classified by their degree of morphological integration into Azerbaijani:
| Class | Description | Example |
|---|---|---|
native |
Pure Azerbaijani, Turkic root | kitab (book), ev (house) |
established_loanword |
Old borrowing, fully absorbed | stol (table, from Russian) |
recent_borrowing |
New borrowing, phonologically adapted | internet, supermarket, kafe |
morphological_hybrid |
Foreign stem + Azerbaijani morphology | postlamaq (EN post + AZ -lamaq) |
code_switch |
Inserted word, no morphological integration | Russian word mid-Azerbaijani sentence |
calque |
Phonological reshaping of foreign word | kompüter (from English computer) |
proper_noun |
Names, place names, brand names | Bakı, Guba, iPhone |
unknown |
Cannot be classified | — |
Languages documented
| Code | Language | Family | Script |
|---|---|---|---|
| az | Azerbaijani | Turkic | Latin (post-1991), Cyrillic (Soviet-era) |
| ru | Russian | Slavic | Cyrillic |
| en | English | Germanic | Latin |
| lez | Lezgi | Northeast Caucasian | Cyrillic |
| tly | Talysh | Iranian | Latin/Cyrillic |
| khv | Khinaliq | Northeast Caucasian (isolate) | Latin |
| ava | Avar | Northeast Caucasian | Cyrillic |
| tkr | Tsakhur | Northeast Caucasian | Cyrillic |
| udi | Udi | Northeast Caucasian | Latin/Georgian |
| ka | Georgian | Kartvelian | Mkhedruli |
| hy | Armenian | Indo-European | Armenian |
| fa | Persian | Iranian | Arabic |
| tr | Turkish | Turkic | Latin |
Key findings
- Code-switching rate: ~43% of entries show code-switching between two or more languages
- Matrix language shift by domain: Azerbaijani dominates commercial and street signage; Russian persists in Soviet-era administrative and technical registers; English dominates technology and social media contexts
- Morphological hybrids: Multiple instances of English/Russian stems combined with Azerbaijani grammatical suffixes (e.g. -lamaq verbalising suffix applied to English social media verbs)
- Script mixing: Several entries show simultaneous use of Latin and Cyrillic scripts on a single sign — a direct artefact of the 1991 script transition
- Geographic gradient: Code-switching rate and language mix vary significantly between Baku city centre, peri-urban suburbs, and village contexts (Khinaliq, Guba)
Known limitations
- Photo OCR accuracy varies with image quality and lighting conditions
- AI-generated fields (OCR, translation, linguistic analysis) have been researcher-reviewed but may contain errors — all AI fields are labelled with source tags
- Dataset skews toward commercial/urban contexts in Baku city centre
- Village data from Khinaliq and Guba is underrepresented relative to linguistic significance
- Audio transcription of spontaneous speech contains approximations
- YouTube transcriptions cover selected excerpts only
Ethical statement
- All photographs taken in public spaces in Azerbaijan
- No personally identifiable information included
- Native speaker data anonymised to initials and age group only
- Audio recordings collected with informed verbal consent recorded on audio
- YouTube data used with explicit written permission from Samantha Parker, Learn Azerbaijani Today (correspondence dated 7 July 2026)
- Dataset released under Creative Commons Attribution 4.0 International (CC BY 4.0)
Citation
If you use this dataset in your research, please cite:
@dataset{manikantan2026baku,
title = {Code-Switching in Post-Soviet Urban Space: A Linguistic
Landscape Dataset from Baku and Northern Azerbaijan},
author = {Manikantan, Sudarshan},
year = {2026},
publisher = {Zenodo},
doi = {10.5281/zenodo.21411905},
url = {https://doi.org/10.5281/zenodo.21411905},
note = {Also available at https://huggingface.co/datasets/sudhirulz/baku-comix},
license = {CC BY 4.0}
}
Replace XXXXXXX with your actual Zenodo DOI after publication.
Acknowledgements
- Learn Azerbaijani Today (Samantha Parker) for permission to use pedagogical audio
- OpenStreetMap contributors for reverse geocoding
- Anthropic Claude API for OCR, linguistic analysis, and word-level annotation
- OpenAI Whisper for audio transcription
- The people of Baku, Guba, Gabala, Khinaliq, and Shamakhi
- Downloads last month
- 40