How to prepare a dataset for publication
When preparing a dataset for publication or archiving, you should make a selection of the material you collected, created, or used during your research.
This selection should be guided by the following principles:
-
Include the data with documentation that underpins the publication to support research integrity and reproducibility.
-
Include materials that are unique or cannot be recreated or recollected easily.
-
Include materials that may be relevant for reuse.
The goal is to provide enough information for others to assess whether your data and related materials can be reused for new research.
Please consult your faculty or institute’s specific guidelines, available on your faculty intranet, and ensure that you comply with all requirements.
What to include in a dataset?
As a good practice, consider including the following:
Documentation
Documentation
-
A description of your dataset in a README file or in the description field of the metadata template.
-
Logbooks, protocols, workflows, procedures, or schemas describing how the data were collected, selected, processed, recoded, or otherwise modified.
-
Survey instruments or questionnaires (these may be in the codebook)
-
Blank consent forms or participant information sheets, where applicable.
-
Additional contextual information that helps others understand and interpret the data.
Research data
Research data
-
The files with de-identified data that underpin your publication.
-
Codebook(s) describing variable names, scales, measurement methods, and coding schemes.
-
Scripts, code, or workflows used to clean, process, analyze, or visualize the data.
Supplementary materials
Supplementary materials
-
Tables, graphs, statistical output files (e.g. SPSS output),, guides, and other supporting materials relevant to understanding or reusing the dataset.
Software
Software
-
The DCC has a dedicated section on how to publish your software FAIR.
In addition, consult your scientific group, colleagues, and supervisors to determine which materials are needed to support the understanding and reuse of the dataset.
File preparation
Before uploading your files to DataverseNL, take a moment to prepare them properly, considering the format of your files, as well as how you structure them if you have many to upload.
Use preferred file formats
Use preferred file formats
Of all the file formats available, it is best to use those that guarantee long-term use without the need for paid licenses or outdated software. Many software programs allow for saving in an open file format.
Before storing your files in these formats:
-
Ensure your data has been de-identified where necessary.
-
Use systematic and consistent file names, especially if your dataset contains many files.
Organize your files
Organize your files
If you have many files to upload to your DataverseNL dataset, they must be organized before uploading.
We recommend that you:
-
Organise your files into a logical and comprehensible folder structure.
-
Use clear and descriptive names for folders and files. Make sure these names are consistent with your publication.
-
Include a README file explaining the contents and organization of the dataset.
Making your files easy to find
DataverseNL displays files in alphanumeric order. To make important files (e.g. README file) appear at the top of the list, consider naming them differently, for example:
-
01_README
-
Documentation
Prepare large datasets
Prepare large datasets
Large datasets with a large number of files require some additional preparation because of the way DataverseNL handles uploaded files:
-
DataverseNL automatically unzips any zipped files.
-
DataverseNL has a limit of handling 1000 files per dataset.
Moreover, it is important to consider the ease with which the readers/new users will experience when consulting your dataset. A dataset that appears too cumbersome to read through will likely put off potential interested researchers.
Guidelines for large datasets
-
Double-zip any zipped files before you upload to end up with a zipped file in DataverseNL.
-
Individual files should not exceed 8 GB in size.
-
Make essential information stand out by adding a number or alphabetical order.
The structure of the folders and files will be visible in the “tree-view” in DataverseNL.