For my Ph.D work I used large volume of data which can not be put into my thesis. So, I am planning to put all this data online, so that other could download freely and cross check my work. It would also help future scholars to do the same work conveniently in future. Providing data online used for research is not commonly practiced in my country. I just want to know is it right(also sensible) to do the same? Won't it lead to any complication in future?

Can you also please tell me what is being practiced in other countries like USA and UK regarding data?

[Note : all data is downloaded from free sources so does not involve any copyright issue]

  • 3
    What do you mean by "free sources", exactly?
    – 410 gone
    Commented Nov 18, 2015 at 8:04
  • 2
    Have you asked your advisor about that? Do you have a hosting space? What discipline are you?
    – Alexandros
    Commented Nov 18, 2015 at 8:34
  • I am from finance background
    – Neeraj
    Commented Nov 18, 2015 at 9:33
  • 5
    Something being freely available for online download does not mean that it’s free to distribute, in particular free to host for yourself. Anyway the corresponding issue would be primarily a legal one to be decided on a per-case basis. Moreover: If those data is freely available, is there any particular reason you want to mirror it?
    – Wrzlprmft
    Commented Nov 18, 2015 at 10:02
  • 1
    No I donot want to just mirror it but in filtered and analyzed way which would be very easy to handle for other.
    – Neeraj
    Commented Nov 18, 2015 at 10:51

2 Answers 2


In general I think this is a good idea, but it might depend on your area. This way others can verify your method and compare it to yours. At least this is how it is used in for example Computer Vision and Machine Learning.

Some examples:

Some theses with a dataset published:

Professor S. Rüger from the Open University UK, wrote in How to write a good PhD thesis and survive the viva (for a PhD in the computing subject):

Some value in a PhD thesis is drawn from careful experimental design. It is best practice to only change one parameter at a time; to use datasets that are publicly available or at least make datasets available; to describe experiments in a way so that they are reproducible; and, particularly in Computing, to set up experiments in an automated batch fashion.

You however have to make sure that you are allowed to (re)publish the data on your web without violating any copyrights (even if it can be downloaded for free), otherwise you can just link to the external data set.

Edit: also see this question about Publicly Available Datasets.


Some things people should consider before making their data freely downloadable. Many of these may not apply to you, but they're general considerations. In some cases these are things that should be decided before data are ever collected.

  1. Are the data truly yours? Does your institution have some claim to the raw data, or to the analysis? Do you have co-authors who also have a claim to the data? Do you have permission from all parties to make this available? Can you document this?
  2. Is there anything confidential in the data?
  3. Is there anything copyrighted in the data? Are there other potential legal concerns about making it available? If you have modified or analyzed the data using software, is the download compatible with the software license?
  4. If humans were involved in any way, is their information anonymized? Did they give permission to make their data available, even in aggregate? Did your institutional review board approve this part of the project? Do you have clear documentation showing this?
  5. Are you willing and able to maintain the data? Will it be on a site that you control and will control for a period of time? If not, who controls it, and are they willing and able to continue to make it available? What is a reasonable time for the data to remain available -- two years? Five? Twenty?

For many studies, 1-4 may not apply, but 5 is something people don't seem to think about very much. Far too often, individual researchers stick their data, or supplementary information or whatever, up on their institutional web sites, and then two years later their IT people do some reorganization and all the links are broken; or the people move to another institution and their pages are all deleted; or a bug hits and no one notices. Or they put the data up on their personal pages, and then GeoCities is bought by Yahoo! and gets shut down in their country. The web five years ago was a very different place from today, and it will be very different in another five years.

One option is Dryad, which promises to store researcher data; see their claims here. I have no experience with them other than downloading data, but the concept seems good.

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