Background: I asked this question to a statistical research audience on CrossValidated but was suggested that here might be more appropriate, so re-posting; keen for any tips!

In spatial analysis (and no doubt, any research field) we often work with a broad range of data and software libraries, increasingly open source. These are all associated with various licence conditions. I feel that my current approaches for tracking data and software provenance and licence agreements should be improved.

Traditional reference management software like Endnote or Zotero is one option, although they seem better geared for managing document references. I have also explored some other approaches such as FOSSA (run by the people behind https://tldrlegal.com/), although this seems specifically geared for software development more than research, and in practice I was unable to get it working.

I predominantly use Python and R for programming analyses; the latter is able to generate a BibTeX citation for libraries and using the function citation(), which is great but it doesn't output licence information (which is viewable from the CRAN url). I am unaware of a similar function for Python (which has similar data available on pypi.org). I could write R or Python functions to do this task, but I don't want to reinvent the wheel! Surely this exists already? But I can't find it.

There are other approaches which sound promising - this Open Science Training Book notes some resources like DataTags (framework for data management, facilitating decisions around data sharing) described in an interesting publication by Sweeney et al. from 2015, but their site datatags.org is no longer accessible. Also from 2015, Nicholas Car from CSIRO appears to have published on their PROMS provenance management system, with code available; this looks like it could be tied in with my Python workflows, although it sounds like it might be challenging for my less technical colleagues to work with (but I will give it a go).

I realise I'm conflating a few things here (reference management for data, software, and management of licence agreements), but these seem quite related and have important implications for how research findings are disseminated.

So, do you have any advice for management of research resources (in particular, data and software libraries, their licence agreements and other provenance metadata) in an environment where we are striving towards 'open science'?

  • 1
    When you distribute software for research, in what form do you distribute and what do you actually distribute? For example, all licenses only take affect once you copy or distribute software. If you write a program using Python and distribute only the source code, you don't actually need to comply with the Python license, for example. That license only takes affect once you distribute Python, too. You might decide to cite Python, though, but citation is a different activity for a different purpose than reading, accepting, and complying with a software license.
    – Brandin
    Nov 28, 2018 at 4:22
  • It's a bit of a broad question, partly because the output can be broad in the research group where I work. We calculate indicators of the built environment, which are calculated from multiple data sources. Its not necessarily analysis code that we would distribute but also the data. I want to make sure I record and understand the licence terms of both software libraries and data to make sure our research outputs (which may be data on online maps, for example) meet the conditions of the resources we use and are in turn appropriately licenced -- and be able to advise my colleagues on this.
    – Carl Higgs
    Nov 28, 2018 at 4:38
  • Here is an example from a CSIRO project (which I am not involved in) which highlights the need for managing disparate licences when delivering derived datasets resulting from an analysis (spatial, or otherwise). As I linked to above, it seems CSIRO (an Australia government research institute) are developing software for provenance management along the lines of what I'm interested in. Given there must be broad need for such data management in projects using OS software and data, I wonder what do others do?
    – Carl Higgs
    Nov 28, 2018 at 5:33


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.