Is there a data dictionary?

The Showcase Website

Metadata and further information about each field can be found in the UK Biobank (UKB) Showcase page for that field. If you Search in Showcase for a field number or a keyword, and select the field page, some of the metadata will be visible. Further information can be found in the various tabs on the field page (Data/Instances/Notes/Related fields/Resources).

See the "Category Notes" for each category that the field is in. The categories are in a hierarchy, and categories higher in the hierarchy may contain important details.   Categories are listed on each field page, under the Data-Field Description.

Some of the Showcase resources with regards to genetics or Olink Proteomics data are also available in the UK Biobank Research Analysis Platform (UKB-RAP), in sub-folders called helper_files.  For example, you can find the Olink Showcase resources by navigating to Bulk/Protein biomarkers/Olink/helper_files

If you want more information on how to navigate through the Showcase website, please see the UK Biobank Showcase User Guide. For a further guide on what types of data are available in UK Biobank, please see this article. 

The UKB-RAP data dictionary

You can produce a data dictionary on the UKB-RAP for your project using the command line interface (CLI). The data dictionary you produce will contain metadata about each of the fields available in your specific project.

Here is a guide on how to do this within the UKB-RAP: 

• Launch a JupyterLab instance within the UKB-RAP. For more information on how to do this, please see this article.
• Make a note of your Project ID and Record ID
  • How to find your Project ID: Select your project in the "Projects" area on the UKB-RAP. Copy the project ID from the right hand panel.
  • How to find your Record ID: Select your Dataset Record item and copy the Record ID from the right hand panel. You may need to select show info panel if this does not automatically appear. 
• Open a terminal in JupyterLab and type dx extract_dataset -ddd project-ID:record-ID  . An example can be found below:

• After a few seconds, three new files should appear on the left. Hover over the .csv files to see which one has a name including “data_dictionary”, and double click to open the file. You will see something like this:

• This data dictionary is not participant data, so it can be downloaded.
• It is worth noting that these files are in your JupyterLab instance storage, not in your main project storage. When the JupyterLab is closed, these files will disappear.
  • Type dx upload app*.csv into your terminal to copy all instance csv files to your main project storage.

For more information about the dx_extract command, please see the index of dx commands page, as well as notebooks such as the A110_Export_participant_data on the UKB GitHub. For more information about getting started with the CLI for the UKB-RAP, please see this page. 

UK Biobank GitHub 

To work with metadata programmatically, there are notebooks available in the UK Biobank GitHub. These notebooks will walk you through how to perform basic operations to access UKB data. For more information about the UKB GitHub, please see the UK Biobank GitHub notebooks article. 

Occasional inconsistencies between the UKB-RAP and Showcase

Please be aware that sometimes there will not be an exact match between the data you will see described on Showcase, and what you will find on the UKB RAP. For example, sometimes there will not be a match between the number of participants in Showcase and the number of participants in the UKB-RAP. This circumstance is explained in this article. 

Another example of this is that sometimes arrayed fields on Showcase have been condensed on the UKB-RAP, causing all the data from the several arrays to be in one comma-separated list on the UKB-RAP. The files in the UKB-RAP “Showcase metadata” folder may also state that these fields are arrayed, which may cause some confusion to researchers. If you wish to double check for this inconsistency, creating a data dictionary for your UKB-RAP project will correctly list whether your files in this field are arrayed.

If you need additional guidance or support please submit a ticket