Manually creating a data dictionary

Video tutorial: Data Dictionaries on the Ag Data Commons

For a spreadsheet

Submit a spreadsheet (.XLS or .XLSX) with one tab for introductory information, and separate data dictionary tabs that correspond to each existing tab in your dataset. For example, if your dataset consists of three tabs, your data dictionary will have four tabs: the first for introductory, background information, and three more to correspond to the three tabs of data. Consider using our data dictionary template to get started.

Best Practices:

  • Submit your data as a spreadsheet or csv

  • One table per tab

  • No extraneous comments

  • No empty cells, columns, or rows (enter n/a if nothing applies)

  • Spell out all abbreviations

  • Element definitions should be stated in the singular, be succinct, and be able to stand alone from other element definitions

If you would rather submit a DOCX or PDF, embed tables in your document so they will be exportable. Submit a .DOCX or .PDF with the following:

  • Introductory and explanatory text

    • Explain context: is if from a singular research article, or a larger project?

    • Provide a URI (Uniform Resource Identifier), which will usually be a URL or a DOI (Digital Object Identifier) for the dataset or related journal article.

    • Other pertinent information such as version, date released, etc.)

  • A listing of elements (fields), in addition to the following:

    • Element source table

    • Element definition

    • Element variables

    • Element data type

    • Element field length

    • Required y/n and/or null value note

For a database

Use this option only if you are unable to automatically generate a machine-readable data dictionary. Guidelines for automatically generating a data dictionary from a database are in the previous section of this document.

We suggest submitting your data dictionary as a spreadsheet (consider using our blank template). If you would rather submit a DOCX or PDF, embed tables in your document so they will be exportable. Submit a DOCX or text searchable PDF with the following:

  • Introductory and explanatory text

    • Explain context: is if from a singular research article, or a larger project?

    • Provide a URI (Uniform Resource Identifier), which will usually be a URL or a DOI (Digital Object Identifier) for the dataset or related journal article.

    • Other pertinent information such as version, date released, etc.)

  • A listing of elements (fields), in addition to the following:

    • Element source table

    • Element definition

    • Element variables

    • Element data type

    • Element field length

    • Required y/n and/or null value note

  • If possible, a data diagram or data model showing the relationships among tables.

For example:

Source: The Pacific Northwest Forest Inventory and Analysis Database