Figshare v. Institutional Repositories.
About twice a week at the moment I’m asked to give advice on something in which I have an internal choice, “use our institutional repository (http://ir.canterbury.ac.nz) or figshare”.
Both have very similar functional outcomes – they provide a permanent online space for material. So, in order to help me, here is an explicit list of the things I think about to make a decision in each case.
| Feature | Figshare | Institutional Repository |
| Truly Permanent URLs | DOIs. Metadata mediated by a 3rd party | Handles – not as well understood as DOIs, but functional. DOIs for unique material being implemented. |
| Findability | Excellent – understood by Google Scholar et al. | Excellent – indexed by local and international organisations. |
| Permanence and Preservation | Intentions are good (actually excellent), but 3rd party commercial entity, so there will always be a concern. | Long term explicit preservation plan, and work undergoing on being archived by National Library. |
| Ease of use | Completely under the control of the end user, excellent interface. | Must be mediated by library staff. Pre-registration of permanent URLs is not simply implemented, but achievable. |
| Content Presentation | Excellent content viewers for tabular data, PDFs, &tc | Simply presents file in a type-agnostic way. becomes end user problem to view & use. |
| Metadata Collection | Simple but effective metadata collection. Other than ORCID, no authority control. | Some control, uses formal metadata schema (Dublic Core, ETD (NDLTD). ORCID is on its way. |
| Metadata Presentation/ Interoperability | Extensive (proprietary) API. https://docs.figshare.com/ and OAI PMH | OAI PMH (‘natch). items exportable as SIP/XML/… |
| Control and Collection | We have no idea what ‘our’ authors are doing | We can collect our institution’s output |
The last item, and I think that is the real sticking point – us Librarians do like to have some control – can be mitigated by buying an institutional subscription to Figshare, and then use the API to transfer items into a dark archive if preservation concerns are so great we think we need to replicate them. Functionally I think they are really similar apart from how content is presented, and what formal metadata is collected – the differences are in the lack of mediation in uploading material (which is a good thing) and the presentation of the content.
I’d be really interested to see what other IR managers and Figshare themselves think. I want the best for the authors, that their work is findable, preserved and described.