Initiative to mobilize talent and partnerships across United Kingdom to coordinate and connect national data science driven research efforts related to COVID-19 to address wider impact of COVID-19 pandemic.National Institute for Health Data Science for England, Wales, Scotland and Northern Ireland, is championing use of health data to respond to COVID-19.

The UK COVID-19 Vocal Audio Dataset is designed for the training and evaluation of machine learning models that classify SARS-CoV-2 infection status or associated respiratory symptoms using vocal audio. The UK Health Security Agency recruited voluntary participants through the national Test and Trace programme and the REACT-1 survey in England from March 2021 to March 2022, during dominant transmission of the Alpha and Delta SARS-CoV-2 variants and some Omicron variant sublineages. Audio recordings of volitional coughs, exhalations, and speech (speech not available in open access version) were collected in the 'Speak up to help beat coronavirus' digital survey alongside demographic, self-reported symptom and respiratory condition data, and linked to SARS-CoV-2 test results. The UK COVID-19 Vocal Audio Dataset represents the largest collection of SARS-CoV-2 PCR-referenced audio recordings to date. PCR results were linked to 70,794 of 72,999 participants and 24,155 of 25,776 positive cases. Respiratory symptoms were reported by 45.62% of participants. This dataset has additional potential uses for bioacoustics research, with 11.30% participants reporting asthma, and 27.20% with linked influenza PCR test results.

Contents

• participant_metadata.csv row-wise, participant identifier indexed information on participant demographics and health status. Please see A large-scale and PCR-referenced vocal audio dataset for COVID-19 for a full description of the dataset.
• audio_metadata.csv row-wise, participant identifier indexed information on three recorded audio modalities, including audio filepaths. Please see A large-scale and PCR-referenced vocal audio dataset for COVID-19 for a full description of the dataset.
• train_test_splits.csv row-wise, participant identifier indexed information on train test splits for the following sets: 'Randomised' train and test set, Standard' train and test set, Matched' train and test sets, 'Longitudinal' test set and 'Matched Longitudinal' test set. Please see Audio-based AI classifiers show no evidence of improved COVID-19 screening over simple symptoms checkers for a full description of the train test splits.
• audio/ directory containing all the recordings in .wav format
  • Due to the large size of the dataset, to assist with ease of download, the audio files have been zipped into covid_data.z{ip, 01-24}. This enables the dataset to be downloaded in short periods, reducing the chances of a dropped internet connection scuppering progress. To unzip, first, ensure that all zip files are in the same directory. Then run the command 'unzip covid_data.zip' or right-click on 'covid_data.zip' and use a programme such as 'The Unarchiver' to open the file.
  • Once extracted, to check the validity of the download, please run the 'python Turing-RSS-Health-Data-Lab-Biomedical-Acoustic-Markers/data-paper/unit-tests.py. All tests should pass with no exceptions. Please clone the GitHub repo detailed below.
• README.md full dataset descriptor.
• DataDictionary_UKCOVID19VocalAudioDataset_OpenAccess.xlsx descriptor of each dataset attribute with the percentage coverage.

Code Base

The accompanying code can be found here: https://github.com/alan-turing-institute/Turing-RSS-Health-Data-Lab-Biomedical-Acoustic-Markers

Citations:

Please cite.

@article{coppock2022,

author = {Coppock, Harry and Nicholson, George and Kiskin, Ivan and Koutra, Vasiliki and Baker, Kieran and Budd, Jobie and Payne, Richard and Karoune, Emma and Hurley, David and Titcomb, Alexander and Egglestone, Sabrina and Cañadas, Ana Tendero and Butler, Lorraine and Jersakova, Radka and Mellor, Jonathon and Patel, Selina and Thornley, Tracey and Diggle, Peter and Richardson, Sylvia and Packham, Josef and Schuller, Björn W. and Pigoli, Davide and Gilmour, Steven and Roberts, Stephen and Holmes, Chris},

title = {Audio-based AI classifiers show no evidence of improved COVID-19 screening over simple symptoms checkers},

journal = {arXiv},

year = {2022},

doi = {10.48550/ARXIV.2212.08570},

url = {https://arxiv.org/abs/2212.08570},

}

@article{budd2022,

author={Jobie Budd and Kieran Baker and Emma Karoune and Harry Coppock and Selina Patel and Ana Tendero Cañadas and Alexander Titcomb and Richard Payne and David Hurley and Sabrina Egglestone and Lorraine Butler and George Nicholson and Ivan Kiskin and Vasiliki Koutra and Radka Jersakova and Peter Diggle and Sylvia Richardson and Bjoern Schuller and Steven Gilmour and Davide Pigoli and Stephen Roberts and Josef Packham Tracey Thornley Chris Holmes},

title={A large-scale and PCR-referenced vocal audio dataset for COVID-19},

year={2022},

journal={arXiv},

doi = {10.48550/ARXIV.2212.07738}

}

@article{Pigoli2022,

author={Davide Pigoli and Kieran Baker and Jobie Budd and Lorraine Butler and Harry Coppock and Sabrina Egglestone and Steven G.\ Gilmour and Chris Holmes and David Hurley and Radka Jersakova and Ivan Kiskin and Vasiliki Koutra and George Nicholson and Joe Packham and Selina Patel and Richard Payne and Stephen J.\ Roberts and Bj\"{o}rn W.\ Schuller and Ana Tendero-Ca$\tilde{n}$adas and Tracey Thornley and Alexander Titcomb},

title={Statistical Design and Analysis for Robust Machine Learning: A Case Study from Covid-19},

year={2022},

journal={arXiv},

doi = {10.48550/ARXIV.2212.08571}

}

The Dublin Core™ Metadata Initiative

- Title: The UK COVID-19 Vocal Audio Dataset, Open Access Edition.

- Creator: The UK Health Security Agency (UKHSA) in collaboration with The Turing-RSS Health Data Lab.

- Subject: COVID-19, Respiratory symptom, Other audio, Cough, Asthma, Influenza.

- Description: The UK COVID-19 Vocal Audio Dataset Open Access Edition is designed for the training and evaluation of machine learning models that classify SARS-CoV-2 infection status or associated respiratory symptoms using vocal audio. The UK Health Security Agency recruited voluntary participants through the national Test and Trace programme and the REACT-1 survey in England from March 2021 to March 2022, during dominant transmission of the Alpha and Delta SARS-CoV-2 variants and some Omicron variant sublineages. Audio recordings of volitional coughs and exhalations were collected in the 'Speak up to help beat coronavirus' digital survey alongside demographic, self-reported symptom and respiratory condition data, and linked to SARS-CoV-2 test results. The UK COVID-19 Vocal Audio Dataset Open Access Edition represents the largest collection of SARS-CoV-2 PCR-referenced audio recordings to date. PCR results were linked to 70,794 of 72,999 participants and 24,155 of 25,776 positive cases. Respiratory symptoms were reported by 45.62% of participants. This dataset has additional potential uses for bioacoustics research, with 11.30% participants reporting asthma, and 27.20% with linked influenza PCR test results.

- Publisher: The UK Health Security Agency (UKHSA).

- Contributor: The UK Health Security Agency (UKHSA) and The Alan Turing Institute.

- Date: 2021-03/2022-03

- Type: Dataset

- Format: Waveform Audio File Format audio/wave, Comma-separated values text/csv

- Identifier: 10.5281/zenodo.10043978

- Source: The UK COVID-19 Vocal Audio Dataset Protected Edition, accessed via application to Accessing UKHSA protected data.

- Language: eng

- Relation: The UK COVID-19 Vocal Audio Dataset Protected Edition, accessed via application to Accessing UKHSA protected data.

- Coverage: United Kingdom, 2021-03/2022-03.

- Rights: Open Government Licence version 3 (OGL v.3), © Crown Copyright UKHSA 2023.

- accessRights: When you use this information under the Open Government Licence, you should include the following attribution: The UK COVID-19 Vocal Audio Dataset Open Access Edition, UK Health Security Agency, 2023, licensed under the Open Government Licence v3.0 and cite the papers detailed above.

This dataset contains daily data trackers for the COVID-19 pandemic, aggregated by month and starting 18.3.20. The first release of COVID-19 data on this platform was on 1.6.20. Updates have been provided on a quarterly basis throughout 2023/24. No updates are currently scheduled for 2024/25 as case rates remain low. The data is accurate as at 8.00 a.m. on 8.4.24. Some narrative for the data covering the latest period is provided here below: Diagnosed cases / episodes • As at 3.4.24 CYC residents have had a total 75,556 covid episodes since the start of the pandemic, a rate of 37,465 per 100,000 of population (using 2021 Mid-Year Population estimates). The cumulative rate in York is similar to the national (37,305) and regional (37,059) averages. • The latest rate of new Covid cases per 100,000 of population for the period 28.3.24 to 3.4.24 in York was 1.49 (3 cases). The national and regional averages at this date were 1.67 and 2.19 respectively (using data published on Gov.uk on 5.4.24).

Links to a range of useful COVID-19 datasets and visualisations. For information and advice in relation to COVID-19 please use the following information from reliable, trusted sources such as the Government, the NHS, Public Health England and the Council. www.gov.uk/coronavirus www.nhs.uk/coronavirus www.calderdale.gov.uk/coronavirus

CVD-COVID-UK, co-ordinated by the British Heart Foundation (BHF) Data Science Centre (https://bhfdatasciencecentre.org/), is one of the NIHR-BHF Cardiovascular Partnership’s National Flagship Projects.

CVD-COVID-UK aims to understand the relationship between COVID-19 and cardiovascular diseases through analyses of de-identified, pseudonymised, linked, nationally collated health datasets across the four nations of the UK. The consortium has over 400 members across more than 50 institutions including data custodians, data scientists and clinicians, all of whom have signed up to an agreed set of principles with an inclusive, open and transparent ethos.

Approved researchers access data within secure trusted/secure research environments (TREs/SDEs) provided by NHS England (England), the National Safe Haven (Scotland), the SAIL Databank (Wales) and the Honest Broker Service (Northern Ireland). A dashboard of datasets available in each nation’s TRE can be found here: https://bhfdatasciencecentre.org/areas/cvd-covid-uk-covid-impact/

This dataset represents the linked datasets for CVD-COVID-UK in the National Safe Haven for Scotland and contains the following datasets: • Outpatient Appointments and Attendances - Scottish Morbidity Record (SMR00) • General Acute Inpatient and Day Case - Scottish Morbidity Record (SMR01) • Scotland Accident and Emergency • COVID-19 Tests (lab/lighthouse testing) • SARS-CoV-2 viral sequencing data (COG-UK data) - Lineage/Variant Data - Scotland • Scottish Covid-19 Vaccination Data • National Records of Scotland (NRS) - Deaths Data • SICSAG Daily (Scottish Intensive Care Audit Group) • SICSAG Episodes (Scottish Intensive Care Audit Group) • Prescribing Information System (PIS) • Scottish Stroke Care Audit • Diabetes covariates • Scottish Renal Registry

Data forming the COVID-19 SARI-Watch data set relate to demographic, risk factor, treatment, and outcome information for patients admitted to hospital with a confirmed COVID-19 diagnosis, as recorded in the PHE COVID-19 SARI-Watch Surveillance System.

SARI-Watch data are to be collected for the purposes of direct care, service monitoring, planning and research in response to the spread of COVID-19, including for the following purposes identified in the COVID-19 Directions (see below): •understanding information about patient access to health services and adult social care services as a direct or indirect result of COVID-19 and the availability and capacity of those services •monitoring and managing the response to COVID-19 by health and social care bodies and the Government, including providing information to the public about COVID-19 and its effectiveness, and information about capacity, medicines, equipment, supplies, services and the workforce within the health services and adult social care services •research and planning in relation to COVID-19, such as providing COVID-19 diagnosis.

Timescales for dissemination can be found under 'Our Service Levels' at the following link: https://digital.nhs.uk/services/data-access-request-service-dars/data-access-request-service-dars-process Standard wording

NHS Digital will only disseminate SARI-Watch data collected from PHE where the information is linked to other information controlled by NHS Digital.

This data set has been designed to support the aim to improve NHS cancer patient outcomes by sharing real-time data about UK cancer services during the COVID-19 pandemic and afterwards. This will be achieved by creating a network of UK-wide hospitals large and representative enough to allow national and local data analysis, enabling insights that are not possible from current national datasets. All hospitals share an agreed minimum dataset at minimum quality on a weekly basis (Level 1) with the option to share deeper datasets if that fits their digital maturity, patient choice, information governance and organisational priorities (Level 2 and 3). All data will be kept and shared under appropriate and agreed information governance and legal contracts, and via Trusted Research Environment(s).

Patient-level data for adults and children with CKD or adults with an AKI on dialysis who are under the care of NHS hospital renal centres in England, Northern Ireland and Wales and who have a positive laboratory test for SARS-CoV-2.

Virus Watch will provide data relevant to a wide range of audiences involved in pandemic response. Virus Watch has collected personal and special category data.

Baseline survey – the baseline survey collects basic demographic information including sex, date of birth, age in years, ethnicity. It also includes details of the household structure, socioeconomic status including household income. The survey also collects health data used for the study including existing medical conditions (general and COVID-related) and access to health during the pandemic.

Weekly survey – the weekly survey collects data about any illnesses within the household during each week and the results of any COVID tests performed. The survey collects information on behaviours during illness of the household. Since Jan 2021, the weekly survey has also collected data on vaccination status of household members.

Monthly surveys – the monthly surveys collect regular data on contact patterns of household throughout the pandemic, regardless of symptoms or illnesses in the household. Each month additional bespoke questions have been asked in the monthly surveys in order to inform important policy questions at the time.

Laboratory data – the laboratory data includes information on the results of antibody tests for a subset of participants including nucleocapsid and spike antibody levels. It also includes PCR results for participants that took part in home COVID testing for Virus Watch.

The current COVID-19 pandemic, caused by the SARS-CoV-2 virus, represents a major threat to health in the UK and globally. To fully understand the transmission and evolution of the virus requires sequencing and analysing viral genomes at scale and speed. The numbers of samples calls for a rapid increase in the UK’s pathogen genome sequencing capacity rapidly and robustly.

To provide this increased capacity to collect, sequence and analyse the whole genomes of virus samples in the UK, the COVID-19 Genomics UK (COG-UK) consortium is pooling the world leading knowledge and expertise in genomics of the four UK Public Health Agencies, multiple regional University hubs, and large sequencing centres such as the Wellcome Sanger Institute.

Samples collected as part of the ONS Covid Infection Survey and sequenced as part of the COG-UK consortium.

Includes: Patient demographics, Source Organisation, vaccination details and vaccine batch events. Its scope covers: Anyone vaccinated within England Anyone vaccinated in a Devoted Administration where this information is subsequently passed to England.

Settings include: hospital hubs - NHS providers vaccinating on site local vaccine services – community or primary care led services which could include primary care facilities, retail, community facilities, temporary structures or roving teams vaccination centres – large sites such as sports and conference venues set up for high volumes of people

Timescales for dissemination can be found under 'Our Service Levels' at the following link: https://digital.nhs.uk/services/data-access-request-service-dars/data-access-request-service-dars-process

This replication archive contains all scripts and data necessary to replicate the analysis in “Risk and Preferences for Government Healthcare Spending: Evidence from the UK COVID-19 Crisis”.

Motivation

Why did I create this dataset? This is my first time creating a notebook in Kaggle and I am interested in learning more about COVID-19 and how different countries are affected by it and why. It might be useful to compare different metrics between different countries. And I also wanted to participate in a challenge, and I've decided to join the COVID-19 datasets challenge. While looking through the projects, I noticed https://www.kaggle.com/koryto/countryinfo and it inspired me to start this project.

Method

My approach is to scour the Internet and Kaggle looking for country data that can potentially have an impact on how the COVID-19 pandemic spreads. In the end, I ended up with the following for each country:

• Monthly temperature and precipitation from Worldbank
• Latitude and longitude
• Population, density, gender and age
• Airport traffic from Worldbank
• COVID-19 date of first case and number of cases and deaths as of March 26, 2020
• 2009 H1N1 flu pandemic cases and deaths obtained from Wikipedia
• Property affordability index and Health care index from Numbeo
• Number of hospital beds and ICU beds from Wikipedia
• Flu and pneumonia death rate from Worldlifeexpectancy.com (Age Adjusted Death Rate Estimates: 2017)
• School closures due to COVID-19
• Number of COVID-19 tests done
• Number of COVID-19 genetic strains
• US Social Distancing Policies from COVID19StatePolicy’s SocialDistancing repository on GitHub
• DHL Global Connectedness Index 2018 (People Breadth scores)
• Datasets have been merged by country name whenever possible. I needed to rename some countries by hand, e.g. US to United Sates, etc. but it's possible that I might have missed some. See the output file covid19_merged.csv for the merged result.

See covid19_data - data_sources.csv for data source details.

Notebook: https://www.kaggle.com/bitsnpieces/covid19-data

Caveats

Since I did not personally collect each datapoint, and because each datasource is different with different objectives, collected at different times, measured in different ways, any inferences from this dataset will need further investigation.

Other interesting sources of information

• IMF Policy Tracker
• nCov strain analysis across the globe!
• Google Mobility Report
• Oxford government response tracker
• Impact on aviation
• Impact on restaurants
• Harvard
• Tableau datahub
• Unemployment outlook from ILO

Acknowledgements

I want to acknowledge the authors of the datasets that made their data publicly available which has made this project possible. Banner image is by Brian.

I hope that the community finds this dataset useful. Feel free to recommend other datasets that you think will be useful / relevant! Thanks for looking.

The dashboard provides data on COVID-19 testing, cases, healthcare, and deaths in the UK. Data are covered by the UK Open Government Licence.

The TwinsUK cohort (https://twinsuk.ac.uk/), set up in 1992, is a major volunteer-based genomic epidemiology resource with longitudinal deep genomic and phenomics data from over 15,000 adult twins (18+) from across the UK who are highly engaged and recallable. The cohort is predominantly female (80%) for historical reasons. It is one of the most deeply characterised adult twin cohort in the world, providing a rich platform for scientists to research health and ageing longitudinally. There are over 700,000 biological samples stored and data collected on twins with repeat measures at multiple timepoints. Extremely large datasets (billions of data points) have been generated for each TwinsUK participant over 30 years, including phenotypes from questionnaires, multiple clinical visits, and record linkage, and genetic and ‘omic data from biological samples. TwinsUK ensures derived datasets from raw data are returned by collaborators to enhance the resource. TwinsUK also holds a wide range of laboratory samples, including plasma, serum, DNA, faecal microbiome and tissue (skin, fat, colonic biopsies) within HTA-regulated facilities at King's College London.

More recently, postal and at-home collection strategies have allowed sample collections from frail twins, our whole cohort for COVID-19 studies, and for new twin recruits. The cohort is recallable either on a four-year longitudinal sweep visit or, based on diagnosis or genotype.

More than 1,000 data access collaborations and 250,000 samples have been shared with external researchers, resulting in over 800 publications since 2012.

TwinsUK is now working to link to twins’ official health, education and environmental records for health research purposes, which will further enhance the resource, education and environmental records for health research purposes, which will further enhance the resource.

COVID-19 UK Non-hospital Antigen Testing Results (Pillar 2) data is required by NHS Digital to support COVID-19 requests for linkage, analysis and dissemination.

Relative odds of respondents having a positive opinion of UK government decision-making during the COVID-19 lockdown, by demographic group.