{"success":true,"data":{"author":"Indian Institute of Science (IISc) Bangalore, Coswara project; 16 kHz Parquet edition on Hugging Face","bucketAccessType":"public","bucketName":"szzs1693/coswara-data","bucketPrefix":"","category":"Health & Biomedical","cloudProvider":"huggingface","createdAt":"2026-11-22T01:09:18.439Z","description":"Coswara, the crowdsourced respiratory sound and speech database collected by the Indian Institute of Science (IISc) Bangalore between April 2020 and February 2022 for COVID-19 screening from audio. Each of 2,746 participants from more than 20 countries (most in India) recorded nine sounds: deep and shallow breathing, heavy and shallow cough, sustained vowels /a/, /e/ and /o/, and counting at normal and fast pace, giving 24,718 audio files, with 681 participants COVID-19 positive. Metadata has 36 fields per participant covering COVID-19 status, demographics, symptoms and comorbidities, plus per-recording audio quality ratings. This Hugging Face edition resamples all audio to 16 kHz WAV and stores it as Parquet in two configurations, metadata and audio, with participant-level train, validation and test splits; load with Hugging Face datasets. Uses include cough and breath sound classification, respiratory disease detection, acoustic biomarkers and health speech analysis. The IISc Coswara-Data repository states the data are licensed under CC BY 4.0. Cite the Coswara data descriptor in Scientific Data, 2023 (doi 10.1038/s41597-023-02266-0) and the Interspeech 2020 Coswara paper (doi 10.21437/Interspeech.2020-2768).","displayTitle":"Coswara: Crowdsourced Breathing, Cough and Voice Recordings for COVID-19 Detection from 2,746 Participants (IISc Bangalore), 16 kHz Parquet","huggingFaceUrl":"https://huggingface.co/datasets/szzs1693/coswara-data","id":"b82dc5e4-864b-5f80-bfea-4a19f7d2ce0b","license":"CC-BY-4.0","licenseEvidenceUrl":"https://raw.githubusercontent.com/iiscleap/Coswara-Data/master/README.md","objectCountEstimate":62,"probeStatus":"ok","referenceUrl":"https://doi.org/10.1038/s41597-023-02266-0","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"szzs1693/coswara-data\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"szzs1693/coswara-data\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=szzs1693, license:cc-by-4.0, size_categories:10K<n<100K, format:parquet, downloads=344"],"format":"mixed","generatedAt":"2026-11-22T01:10:17.107Z","layout":{"example":"audio/test/audio-test-00001.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"audio/test/audio-test-00001.parquet","truncatedAt":512}],"sampleObjects":["audio/test/audio-test-00001.parquet","audio/test/audio-test-00002.parquet","audio/test/audio-test-00003.parquet","audio/test/audio-test-00004.parquet","audio/test/audio-test-00005.parquet"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\ntask_categories:\n- audio-classification\ntags:\n- covid-19\n- health\n- audio\n- medical-diagnosis\n- speech\n- breathing\n- cough\nsize_categories:\n- 10K<n<100K\nlanguage:\n- en\npretty_name: Coswara COVID-19 Audio Dataset\nconfigs:\n- config_name: audio\n  data_files:\n  - split: train\n    path: audio/train/*.parquet\n  - split: val\n    path: audio/val/*.parquet\n  - split: test\n    path: audio/test/*.parquet\n- config_name: metadata\n  data_files:\n  - split: train\n    path: metadata/train.parquet\n  - split: val\n    path: metadata/val.parquet\n  - split: test\n    path: metadata/test.parquet\n---\n\n# Coswara COVID-19 Audio Dataset\n\nThis is the derived dataset of [Coswara Project Dataset](https://github.com/iiscleap/Coswara-Data), with audio normalized by downsampled to 16 kHz for storage efficiency. The dataset has been converted to Parquet format, and hosted on Hugging Face for more accessibility.\n\n## Dataset Description\n\nThe Coswara dataset is a crowdsourced collection of respiratory sounds and speech recordings for COVID-19 diagnosis research, collected by the Indian Institute of Science (IISc) Bangalore. The dataset contains audio recordings of breathing, cough, and speech from 2,746 participants collected between April 2020 and February 2022.\n\n### Dataset Summary\n\n- **Total Participants:** 2,746\n- **Total Audio Files:** 24,718 (9 recordings per participant)\n- **Collection Period:** April 2020 - February 2022\n- **Audio Format:** 16kHz WAV (resampled from mixed source rates: 48kHz, 44.1kHz, 16kHz)\n- **Geographic Coverage:** 20+ countries (91.6% India)\n- **COVID-19 Positive:** 681 participants (24.8%)\n\n### Supported Tasks\n\n- **Audio Classification:** COVID-19 detection from respiratory sounds\n- **Multi-task Learning:** Joint prediction using multiple audio types\n- **Health Status Classification:** 8-class health status prediction\n- **Symptom Analysis:** Correlation between audio features and COVID symptoms\n\n## Dataset Structure\n\n### Configurations\n\nThe dataset provides two configurations (subsets):\n\n#### 1. `metadata` Configuration\n\nParticipant health and demographic information.\n\n**Splits:**\n\n- `train`: 1,922 participants\n- `val`: 412 participants\n- `test`: 412 participants\n\n**Features:** 36 fields including COVID status, demographics, symptoms, and comorbidities\n\n#### 2. `audio` Configuration\n\nIndividual audio recordings with quality annotations and denormalized metadata.\n\n**Splits:**\n\n- `train`: ~17,298 audio files (1,922 participants × 9)\n- `val`: ~3,708 audio files (412 participants × 9)\n- `test`: ~3,708 audio files (412 participants × 9)\n\n**Features:** Audio arrays (16kHz) + quality scores + 10 key metadata fields\n\n### Audio Types\n\nEach participant provided 9 recordings:\n\n1. **breathing-deep** - Deep breathing sounds\n2. **breathing-shallow** - Shallow breathing sounds\n3. **cough-heavy** - Heavy cough sounds\n4. **cough-shallow** - Shallow cough sounds\n5. **vowel-a** - Vowel 'a' pronunciation\n6. **vowel-e** - Vowel 'e' pronunciation\n7. **vowel-o** - Vowel 'o' pronunciation\n8. **counting-normal** - Counting at normal speed\n9. **counting-fast** - Counting at fast speed\n\n## Schema\n\n### `metadata` Configuration\n\n| Field                       | Type   | Description                                                                  |\n| --------------------------- | ------ | ---------------------------------------------------------------------------- |\n| `id`                        | string | Unique participant identifier (hash)                                         |\n| `a`                         | int64  | Age in years                                                                 |\n| `g`                         | string | Gender (male/female/other)                                                   |\n| `covid_status`              | string | Health status (8 categories, see below)                                      |\n| `record_date`               | date   | Recording submission date                                                    |\n| `test_status`               | string | COVID test result (p=positive, n=negati","encoding":"utf8","key":"README.md","name":"README.md","size":19557,"truncated":true}],"sourceObjectKey":"audio/test/audio-test-00001.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"szzs1693/coswara-data\")\nprint(ds)","starterSamplePath":"audio/test/audio-test-00001.parquet","totalObjects":62},"summary":"Coswara, the crowdsourced respiratory sound and speech database collected by the Indian Institute of Science (IISc) Bangalore between April 2020 and February 2022 for COVID-19…","title":"Coswara: Crowdsourced Breathing, Cough and Voice Recordings for COVID-19 Detection from 2,746 Participants (IISc Bangalore), 16 kHz Parquet","updatedAt":"2026-11-22T03:35:51.927Z","visibility":"public","websiteUrl":"https://github.com/iiscleap/Coswara-Data"}}