Metadata Best Practices#

You are here

User Guide → Metadata Best Practices

This page provides tips for high-quality metadata that supports FAIR principles. If you landed here from a search, you may want to start with the User Guide overview or Getting Started for context on the contributor role.

High-quality metadata is the foundation of FAIR (Findable, Accessible, Interoperable, Reusable) data. This guide provides practical tips for ensuring your data contributions are as valuable as possible for future research.

Why Metadata Matters#

When you contribute data to a FairDM portal, you’re not just recording measurements—you’re creating a lasting scientific record. Complete, accurate metadata ensures that:

  • Future researchers (including your future self) can understand and reuse your data

  • Collaborators can find relevant datasets and samples quickly

  • Data repositories can correctly index and preserve your work

  • Funders and institutions can demonstrate research impact and compliance

Important

FAIR Principle: “Data should be as open as possible, and as closed as necessary—but metadata should always be open.” Even if data access is restricted, good metadata makes your research discoverable.

Required vs. Optional Fields#

Always Provide Required Fields#

Required fields (marked with a red asterisk *) are essential for identifying and locating samples and measurements. These typically include:

  • Sample ID: A unique identifier following your lab or field convention

  • Collection Date: When the sample was collected

  • Collection Location: Geographic coordinates or site name

  • Sample Type: Rock, water, sediment, etc.

Why they matter: Without these basics, samples cannot be reliably identified or located in the future.

Optional Fields Are Still Valuable#

Don’t skip optional fields just because they’re optional:

  • Notes: Additional observations or anomalies

  • Analyst: Who performed the analysis

  • Environmental Conditions: Temperature, weather, etc., during sampling

  • References: Related publications or protocols

Why they matter: These details often become critical when data is reused years later.

Tip

When in doubt, include it: If you’re unsure whether to record a piece of information, err on the side of including it. You can’t go back in time to collect metadata you didn’t record.

Use Controlled Vocabularies When Available#

Many fields in FairDM portals use controlled vocabularies—predefined lists of standardized terms.

Why Use Controlled Vocabularies?#

  • Consistency: Everyone uses the same term for the same concept (e.g., “pH” instead of “acidity”)

  • Searchability: Users can find all measurements of a specific type

  • Interoperability: Data can be compared across projects and institutions

How to Use Them#

When you encounter a dropdown field:

  1. Select from the list whenever possible rather than entering custom text

  2. Check existing entries in the dataset to see which terms are already in use

  3. Contact your administrator if the vocabulary is missing a term you need—they can add it system-wide

Example: If your portal has a controlled vocabulary for rock types, always select from the dropdown rather than typing “granite” manually in a free-text field.

Record Provenance and Methods#

Provenance is the history of a sample or measurement—where it came from, who collected it, how it was processed, and how it was analyzed.

Best Practices for Provenance#

For Samples#

  • Record the collector: If known, note who collected the sample

  • Document sample preparation: If the sample was processed (e.g., filtered, dried, ground), describe the steps

  • Link to related samples: If this sample is a subsample or related to another sample, note the connection

For Measurements#

  • Specify the method: Don’t just say “pH meter”—include the make/model and calibration details

  • Record the analyst: Who performed the analysis?

  • Document calibration and QA/QC: Include calibration standards, blanks, replicates, etc.

  • Note instrument settings: If relevant, record settings like detection limits, wavelengths, etc.

Example (Good Provenance): Method: Water pH measured using Hach HQ40d handheld meter (serial #12345), calibrated with NIST-traceable pH 4.0, 7.0, and 10.0 buffers immediately prior to measurement. Measurement precision ±0.05 pH units based on replicate measurements (n=3) of field blank.

Example (Insufficient Provenance): Method: pH meter

Important

Why it matters: Future researchers need to know if measurements from different campaigns are directly comparable, or if method differences require calibration.

Be Specific and Consistent#

Use Specific, Descriptive Names#

  • Good: “River water sample from Site B main channel, low-flow conditions, June 2024”

  • Poor: “Water sample 45”

Maintain Consistent Naming Conventions#

If your dataset uses a naming pattern (e.g., “SITE-YEAR-SEQUENCE”), continue it:

  • Good: SITE-B-2024-045, SITE-B-2024-046, SITE-B-2024-047

  • Poor: SITE-B-2024-045, Site B 46, B-047

Use Standard Units and Formats#

  • Dates: Use ISO 8601 format (YYYY-MM-DD) whenever possible

  • Coordinates: Decimal degrees (e.g., 45.5231, -122.6765) are preferred over degrees/minutes/seconds

  • Units: Use standard abbreviations (e.g., mg/L, °C, ppm) and always include them with numeric values

Tip

Coordinate consistency: If your portal uses a specific coordinate reference system (e.g., WGS84), always provide coordinates in that system. Don’t mix coordinate systems within a single dataset.

Review Before Saving#

Before clicking “Save,” take a moment to:

  1. Check for typos: Especially in sample IDs and numeric values

  2. Verify required fields: Are they all filled in?

  3. Ensure units are specified: For all numeric measurements

  4. Review free-text fields: Are descriptions clear and complete?

  5. Confirm links: If you linked to a related sample, project, or contributor, is it correct?

Tip

Use the preview: If your portal offers a preview mode, use it to see how your entry will appear to others.

Common Pitfalls to Avoid#

❌ Don’t leave fields blank without reason: If a field is optional but you have the information, provide it.

❌ Don’t use ambiguous abbreviations: What’s obvious to you now may be cryptic in five years (or to other researchers).

❌ Don’t mix units: If most measurements are in mg/L, don’t switch to ppm without clear documentation.

❌ Don’t skip methods: “Standard method” or “usual procedure” is not sufficient. Reference specific protocols or SOPs.

❌ Don’t enter placeholder data: If you don’t have a value yet, leave the field blank rather than entering “TBD” or “999.”

When to Seek Help#

Contact your portal administrator if:

  • A required field doesn’t make sense for your sample type

  • The controlled vocabulary is missing a critical term

  • You’re unsure how to record a complex measurement or relationship

  • You need to correct or delete data after submission

Additional Resources#

Note

Keep learning: Metadata best practices evolve. Check back periodically for updates to these guidelines, and share feedback with your portal team to improve documentation for future contributors.