How to write metadata others can reuse#

Metadata is what lets someone who has never met you judge a dataset, find it and cite it. Follow these steps when you fill in a dataset, and again before you ask for it to be published. The reasons are in Why complete metadata matters.

Before you start#

You need the Can edit level on the dataset. Create the dataset first if it does not exist yet.

Steps#

  1. Name the dataset for what it holds, where and when, for example “River water chemistry, Site B, 2024”. Avoid a bare abbreviation or a number. Enter it in Name on Edit details.

  2. Write the abstract. Open Edit descriptions in the Manage menu and fill in the Abstract area with what the data is, why it was collected and what it covers. Write it so that someone outside your field can tell whether the data is useful to them.

  3. Write the methods in the Methods area of the same screen. Name each instrument by make and model, and give its calibration, the processing steps and the software you used. “pH meter” tells a reader nothing. “Hach HQ40d handheld meter, calibrated with pH 4.0, 7.0 and 10.0 buffers before each measurement” tells them whether their data are comparable. See Descriptions.

  4. Give units wherever a description states a number. Use standard abbreviations such as mg/L and °C, and use the same unit throughout. Do not switch between units without saying so.

  5. Record the collection period. Open Key dates and enter the Collection Start Date and Collection End Date. Enter each date as precisely as you know it, as a year, a month or a day. See Key dates.

  6. Add keywords that people would search for. Open Keywords and type them. Use the same spelling every time. See Keywords.

  7. Record the dataset’s DOI if it has one. Open Identifiers and choose DOI as the type. See Identifiers.

  8. Check the licence. On Edit details, make sure License is the one you intend, because it says what others may do with the data. If a paper describes the dataset, choose it in Data Publication.

  9. Credit everyone who made the dataset, and choose a role for each. Name a contact person, because the readiness card asks for one. Put people in the order the citation should name them. Only a manager can change the list. See Contributors.

  10. Open the dataset page and read the Ready to publish? card. Each missing required item has a Fix link. Work through them until the card says “Everything required is in place.”

  11. Read each screen again before you save it. Check names and numbers for typos, check that every number has its unit and check that a reader would understand each description without asking you.

The result#

The dataset page shows a complete description, the collection period on its Timeline, keywords in the header, the people who made it and a citation that names them. The recommended items on the readiness card, such as keywords and a related publication, are in place as well.

Leave a field empty when you do not have the value yet. A placeholder such as “TBD” or “999” reads as data to the next person.