Splash image

Research data support at the Bolin Centre

Publish data at the Bolin Centre Database

Table of Contents

  1. How to publish data
    1. Sign in
    2. See your datasets
    3. Create new dataset
    4. Revise metadata
    5. Get support
    6. Upload data files
    7. Share draft
      1. Open link with only read access
      2. Personal invitation with update access
    8. Data curation
    9. DOI
    10. Publish dataset
    11. Publish new version
    12. Edit metadata after publication
    13. Examples
    14. Large data
    15. Data size reduction
      1. Observations
      2. Processed data
      3. Model simulations
      4. Sub-set for publication
      5. Temporary extensive model output
  2. How to publish source code

How to publish data

Sign in

  1. Go to the Bolin Centre Database main page.
  2. Click on 'Contribute' in the upper right corner.
  3. Click on 'Sign in' (blue button).
  4. Click on 'Sign in with Orcid' and sign in using your Orcid credentials.

See your datasets

  1. When you have signed in, a list with all your datasets is loaded. If you don't have any datasets yet, the list is empty.
  2. If you already have some datasets, the list will by default only show those that are not yet public.
  3. To see all your datasets, including those that are public, click 'Show all'.

Create new dataset

  1. Click 'New dataset' (blue button).
  2. Use our data contribution form and enter the information you want.
  3. Tip 1: Click on 'preview' in the upper left of the form. Then, you will see all entered text to the right. The preview will be updated interactively when you type anything in the form.
  4. Tip 2: Click 'More info' at each metadata field to learn more about what kind of information to fill in. In some fields there is also a second 'More info' for further details.
  5. When you have done all changes you want, click the blue 'Submit' button at the bottom of the entry form.
  6. Click 'Continue'.

Revise metadata

The information that you provide about your dataset is called metadata.

  1. In your dataset list, click on (or open in a new tab) the dataset that you want to revise.
  2. You should now see a presentation of the latest revision.
  3. To make a new revision, click on the blue 'Update' button.
  4. Use our data contribution form and make the changes you want.
  5. Tip 1: Click on 'preview' in the upper left of the form. Then, you will see all metadata to the right. The preview will be updated interactively when you type anything in the form.
  6. Tip 2: Click 'More info' at each metadata field to learn more about what kind of information to fill in. In some fields there is also a second 'More info' for further details.
  7. When you have done all changes you want, click the blue 'Submit' button.
  8. Click 'Continue' and make further changes by repeating steps 3 to 7 until you are satisfied.

Get support

Please do not hesitate to contact us at bolindata@su.se if you have questions or need assistance with filling in the Data contribution form.

Upload data files

  1. You can upload your data file in the section 'Data' in our Data contribution form.
  2. Note: Before you can upload data, the database staff must first upgrade the dataset's status. When needed, send an email to bolindata@su.se so that we can help you to proceed.
  3. Upload data when you have got the permission.
  4. It is only possible to upload one file. If your dataset includes more files, put them in one zip file and upload it.
  5. The default maximum size for upload is 10 GB. The database staff can give you permission (which is valid for 24 hours) to upload more. If this is needed, send an email to bolindata@su.se with information of the total file size that you wish to upload.

Share draft

You can share your dataset with others before it has been published. You can choose between giving only read access (e.g. for journal reviewers) or giving update access (e.g. your near collaborators).

  1. In your dataset list, click on the dataset that you want to share.
  2. Click on 'Share' (blue button).
  3. Copy the web link provided for read-only access.
  4. Send the link to the persons you want. Anyone with the link can see entered information and access the data file for one particular revision, but not for any other revisions.
Personal invitation with update access
  1. In your dataset list, click on the dataset that you want to share.
  2. Click on 'Share' (blue button).
  3. Type the ORCiD ID for the person you want to share the dataset with. To find this, search for the person's name on the ORCiD website. Just type the actual ID string, which has the format 0000-1234-5678-9101. Do not enter the full web address!
  4. Click on 'Share' (green button).
  5. Re-load your browser tab. You should then see the entered ID string in the list of invited users.
  6. If you want to invite more persons, repeat steps 3 to 5.

Data curation

  1. Send us an email to bolindata@su.se when you have entered as much information as you can on your own, and give us information about when you want the dataset to be published.
  2. We will review your data and metadata. We will give feeback on how it can be improved, in order to fulfil general requirements for data publishing.
  3. This is called data curation. It is an iterative process between you and us, which can be quick (a day) or slow (weeks) depending on the complexity.

DOI

  1. Each dataset in the Bolin Centre Database is assigned a unique DOI. This makes it easy to cite and find your data. If a dataset has more than one version, each version will have its own DOI. The DOI will be created when the dataset is published.
  2. A preliminary DOI string will be automatically generated when you submit your first revision of the draft dataset. This will not work! You should not forward this preliminary string to anybody.
  3. We will decide the final DOI after we have understood your data topic. This decision might involve some correspondence.
  4. You can provide the final DOI to a journal when we have decided what the DOI will be.
  5. Note: No real DOI will exist until the dataset has been published. Until then, it is only a preliminary text string that can be changed.

Publish dataset

  1. When you and we agree that the dataset and its metadata is ready to be published, the database staff will approve it.
  2. You can then publish your dataset yourself whenever it suits you, but first you must sign an agreement.
  3. In your dataset list, click on the dataset that you want to publish.
  4. Click on 'Agreement' (button), read through the text before you sign.
  5. Click on 'Publish' (green button).
  6. Your dataset will then be visible to the public and searchable on the Bolin Centre Database main page.
  7. The DOI is automatically created and made findable at the same time as the dataset is published.

Publish new version

  1. To publish a new version of a dataset, follow the instructions under Revise metadata.
  2. Revise your metadata, including a new version number and an explanation of the difference compared to the previous version.
  3. Follow instructions under Upload data files.
  4. Follow instructions under Publish dataset.

Edit metadata after publication

  1. You can make minor changes in metadata for a public version of a dataset, without any need to publish a new version and creating a new DOI. For example, you can add a reference to an article that describes the dataset.
  2. Follow the instructions under Revise metadata.
  3. Make your minor revision of the metadata, but do not change the version number.
  4. Follow instructions 1 to 6 under Publish dataset.

Examples

We recommend that you look at some recently published datasets to see different examples of how metadata can be written. Our latest datasets are always found at the top of the dataset list on our start page.

Large data

If your data is very large (> 10 TB) and it is more appropriate to store it permanently somewhere else than in the Bolin Centre Database, your metadata should include an internet address leading to the data.

Data size reduction

Some research projects can result in very large datasets. All data may, or can, not always be stored forever. It is therefore necessary to reduce the data size before they are published and archived. For datasets larger than 200 GB we recommend to reduce the size before publication. However, for datasets that are considered particularly important we can accept substantially larger data volumes, in the order of TB.

Primarily, we ask you to keep the data values but reduce the file sizes by:

  • Compressing data
  • Reducing data precision
Observations

Publish everything or as much as possible, including raw data.

Processed data

Publish entire large data products such as interpolated grids, re-analyses, and other important large datasets which are highly requested.

Model simulations

We divide model data into two types:

  • Permanent small sub-set for publication, where all results may not be fully re-produced
  • Temporary extensive model output for the scientific review process, where results can be re-produced
Sub-set for publication

Most simulation data, except CMIP and similar, is probably irrelevant for future scientists to re-use. Therefore, we publish only a small sub-set. However, we encourage scientists to sub-set a bit more than only “figure data” but try very hard to get under 200 GB. This may imply that the results can not be fully re-produced.

We strongly urge scientists to provide source code, model configurations and setup at our code repository.

Some ways to reduce the data size:

  1. Regional subset
  2. Parameters subset
  3. Temporal subset
  4. Mixture of complete data for some variables, but reduced for other
  5. Figure data
  6. Summary statistics of raw data (processed data)
  7. Compress data
Temporary extensive model output

If desired, you can provide temporary extensive model output, generously sub-setted, with sufficient metadata, for the scientific review process, which will be deleted after one year.

How to publish source code

Browse to our code repository, register and upload your code. You can use the web interface or a command line interface, whichever you prefer.

Please refer to our Instructions for publishing source code.