Research Data

Review academic literature, where many published studies generate and analyze data. Search for articles related to the population, problem, or methodology you are targeting. These articles should reference the data repositories, databases, or specific datasets used to conduct the research, and may detail where data has been deposited.

• Select data-specific search filters. For example, in PubMed, you can search articles with Associated Data by selecting this filter under Article Attributes in the sidebar. You can also select Dataset as an Article Type. If Dataset doesn’t appear under Article Type in your sidebar, click Additional Filters button and add it.

• Browse articles for sections such as Supplementary Materials, Data Availability Statement, or Data Citations to find attached data files and links to external data. 

• Use data as a search term. For example, “nonalcoholic fatty liver disease AND data”.

• Enter your search terms in Yale Library’s Books+ and then filter by format Data Sets.

• Search across all National Center for Biotechnology Information (NCBI) resources to see many National Library of Medicine resources at once, including literature and databases.

Find storage locations for datasets and information about those datasets (such as metadata and documentation):

• Use a data repository registry, such as re3data.org and fairsharing.org

• Find data repositories by funder, such as the National Institutes of Health (NIH).

• Find data repositories by discipline, such as those recommended by Nature’s Scientific Data and PLoS ONE.

• Search generalist data repositories, such as Dataverse, Dryad, Figshare, Inter-University Consortium for Political and Social Research (ICPSR), Mendeley Data, Open Science Framework, Qualitative Data Repository (QDR), Synapse, Vivli, and Zenodo.

• Use a government data portal, such as data.gov and healthdata.gov.

Find data that has been purchased by the library or departments, data generated by your colleagues, and data available through some other means. Data at Yale that may be of interest to health sciences researchers includes:

• Yale University Open Data Access (YODA) Project, available from Center for Outcomes Research and Evaluation

• American Hospital Association (AHA), available from Yale Library on WRDS

• Merative MarketScan Database (must be on Yale VPN to access the link), available from Yale Biomedical Informatics and Computing (YBIC) and Biomedical Informatics & Data Science (BIDS)

Consider who might be creating the data you need for research. This includes governing bodies, academic and research institutions, biomedical companies, and not-for-profit organizations, as well as news organizations and other invested groups.

Examples of these include:

• New York Times’ COVID-19 data on Github (read more about this data)

• World Health Organization’s data resources

• United Nations Refugee Agency data catalog

• Gapminder data (read more about Gapminder)

Some academic journals focus on publishing data, and information about data science. Try searching one of the following: 

• GigaScience: publishes all research objects from big data studies across the entire spectrum of life and biomedical sciences.

• Scientific Data: a peer-reviewed, open-access journal for descriptions of scientifically valuable datasets and research that advances the sharing and reuse of scientific data.

• Data in Brief: a multidisciplinary, open access, peer-reviewed journal, publishing short articles that describe and provide access to research data.

• Journal of Open Psychology Data: features peer-reviewed data papers describing psychology datasets with high reuse potential.

• Open Health Data: features peer-reviewed data papers describing health datasets with high reuse potential.

Data reuse is the final R of FAIR, a best practice standard for data sharing. FAIR stands for findable, accessible, interoperable, and reusable. Clear descriptions are at the heart of data reuse. When reusing data, look for well-described data, where the data’s context is apparent. When publishing data, ensure you document it thoroughly, well enough that others can understand and reuse it, and if possible and appropriate, consider publishing your data openly.

Open data is data that can be freely used, re-used, and redistributed by and to anyone as publicly available resources — definition adapted from Open Knowledge Foundation.

Reuse data to:

• Verify your own research.

• Mine the data for new insights.

• Work with data that matches a population or problem you’re interested in — without the overhead of data collection or generation.

• Increase collaboration by analyzing someone else’s data and connecting with like-minded data producers.

• Make open science a standard practice in medicine.

Make your data reusable to:

• Comply with funder mandates — such as NIH’s and NSF’s — and scientific transparency standards.

• Allow others to verify and validate your findings, and potentially collaborate with you.

• Propagate the research cycle and fuel new discoveries, by allowing someone to derive new findings from your data.

• Contribute to the process of tracking scientific inquiry over time.

• Allow citizen scientists to view and interact with health sciences data about their own conditions.

• Align with open science aims as set forth by many professional and cultural organizations, including the UN, UNESCO, and the National Academies of Science, Engineering, and Medicine

• Licensing. Some data are licensed under certain terms — a common one is that you won’t attempt to re-identify research subjects — and some data require you to sign a data use agreement. Read licenses and other agreements/terms carefully, and ensure you and your research team can comply. 

• Access. Sometimes, you have to fill out a data request form, or contact the creator(s) directly and ask for data access. This process can take time.

• Lack of context. As noted above in ‘what’s data reuse?,’ data documentation is central to whether data can be reused. If you’re having difficulty understanding what data variables mean, or how the data was produced, you may not be able to reuse the data.

• Technical difficulties. Sometimes, technical difficulties, such as not enough data storage or unfamiliar data formats, prevent you from accessing a dataset. Reach out to the Medical Library for help before deciding if this is a barrier, though.

• Fees. Not all data are free. If the dataset is not yet in Yale’s collections, consider requesting we purchase it through this form. Additionally, sometimes similar datasets can be found for free. Consider consulting our “Find Data” page.