CyVerse logo

Home_Icon Learning Center Home

2. Organize data

There are several steps to properly organizing your dataset. These include determining what data to include, how many identifiers to request, how to organize the data into folders, and creating the ReadMe file and data inventory.

2.1. Determine what to include 2.2. Determine how many permanent identifiers to request 2.3. Organize your data into folder(s) 2.4. Name your top level folder according to the guidelines 2.5. Create a ReadMe file 2.6. Create an inventory


2.1. Determine what to include:

  1. A data collection may be composed of multiple files and different datasets. In preparing your data for publication:

2. Identify the data and other materials that you consider useful for validation and reuse of your research: - Data associated to a research project may include multiple files with different roles. - If there are components of your dataset that belong in a public repository such as NCBI (e.g., fastq files), submit them to the repository, rather than to CyVerse Curated Data. - Beyond data, you will include the ReadMe file (see Step 2.5), and you may include scripts or links to scripts to run your analysis. Links to analysis tools can also be included as metadata (see next step).


2.2. Determine how many permanent identifiers to request:

To determine how many DOIs to request for a given data collection, consider the following:

  • Think about its size and components.
  • How many studies or publications does it represent?
  • Is your data collection formed by different datasets and are those likely to be used separately?
  • Do you want to create a data collection with one DOI for the entire project and additional related DOIs for distinct datasets so that they are cited individually? DOIs can be nested, so that one dataset is part of another.
  • If you are uncertain about how many DOIs to request, contact us at doi@cyverse.org.

2.3. Organize your data into folder(s):

  1. Organize your data so that there is one folder for each DOI, named according to step 2.4.
  2. Within a folder, include all files in your data package plus the ReadMe file and the inventory.
    • You may have subfolders within a data package.
    • You may include compressed files in a package, as described on the Permanent Identifier FAQs, but do not compress the entire folder/package.

2.4. Name your top level folder according to the guidelines:

The folder containing your dataset should be named using the $Creator_$subject_$date format.

For more details on folder naming, see the CyVerse Curated Data Folder-Naming Guidelines.


2.5. Create a ReadMe file:

Create a text file labeled “readMe” with the following information:


2.6. Create an inventory:

You must create a plain text document that includes an inventory of the contents of the organized dataset (at a minimum, your dataset will contain one data file and one ReadMe file).

The inventory may be part of the ReadMe file or a separate file.

If separate, the inventory should include the ReadMe file and any other additional non-data materials you add to your dataset.

If your dataset contains folders with many files (e.g., large collections of images), you do not need to list each file in the inventory. Simply describe the folder and what it contains.


Additional information, help

The UPenn library offers this guide to organizing data.

DataONE best practices for document storage systems.

Search for an answer: CyVerse Learning Center or CyVerse Wiki

Post your question to the user forum: Ask CyVerse


Home_Icon Learning Center Home