2

(I have seen multiple questions about how to share datasets in general, but none of them seems to involve the criterion of an open contributorship.)

I want to share a dataset to which anyone should be able to contribute.

What would be a suitable platform for such purposes?

The platform should be:

  • free
  • use open licenses (e.g., Creative Commons)
  • render datasets citable (e.g., perhaps with a DOI & related metadata)
  • allow others to contribute in a trivially easy way (e.g., a Git-based system might be too difficult for many to use)
  • guarantee mid-term longevity (e.g., not based on primarily commercial considerations)
  • every contribution should be made transparent (similar to GitHub)

Thus, the following are not suitable:

  • GitHub
  • ResearchGate
  • Academia.edu
  • Google Sheets

Now it seems that there is an overhwelming plurality of platforms that might be suitable for my purposes, such as Zenodo, Figshare, Dataverse, Dryad, OSF, institutional repositories - - but it is difficult to discern their features and to get a good overview if I haven't used any of them yet. May I thus ask you for your experiences or recommendations?

Improve this question
7
  • 4
    I would worry about compatible formats, possibly leading to chaos. I would worry about sabotage. There are issues without a gatekeeper.
    – Buffy
    Commented Mar 16, 2022 at 19:10
  • What kind of data? I would look for a topic-specific platform.
    – Louic
    Commented Mar 16, 2022 at 19:20
  • 4
    You can only trust data contributed by the public as much as you trust the public... for which you must consider who the public consists of and all the potential actors with goals contrary to your own.
    – Bryan Krause ♦
    Commented Mar 16, 2022 at 19:38
  • 1
    I don't understand why you reject Google Sheets? I know of a fair number of community projects on G Sheets that are open (e.g. lists of resources, data)
    – Azor Ahai -him-
    Commented Mar 16, 2022 at 19:39
  • 3
    @nick012000 They definitely have revision tracking, although I suppose you'd have to log in to attribute it
    – Azor Ahai -him-
    Commented Mar 16, 2022 at 22:09

2 Answers 2

Reset to default
2

If you want these two criteria:

allow others to contribute (and access) in a trivially easy way

every contribution should be made transparent

You need to build your own website with a custom database that suits your contributors and data quality process. Anything else will not be both transparent and easy.

Examples: https://icsd.products.fiz-karlsruhe.de/ https://oeis.org/

If you do not care about ease of reading and writing the database, then Git is the solution you want. I would not assume any single Git hosting service will have longevity, but Git does not require you to use a host.

Improve this answer
0

My suggestion: use a simple wiki for the part where people submit contributions to the dataset. There are many systems to choose from. It can be hosted on your academic website or some external provider. You (and optionally a few trusted people) can administrate it. This is important otherwise your dataset risks becoming a mess, and people will quickly stop contributing. In particular the administrators should take care of:

  • Preventing any spam or irrelevant content,
  • Preventing any damage/deletion,
  • Ensuring the quality of the data submitted and the consistency of the format,
  • Periodically releasing an official new version of the full dataset

Most wikis are easy to use, and you can have help pages explaining how to contribute. The regular release of the official dataset can be made available on one of the sites that you mention.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .