Fig share
| 2 March, 2011 | Richard P. Grant |
2
|
|
No, not something to do with fruit. We’ve been thinking a bit about open data recently, talking to Peter Murray-Rust and Cameron Neylon, and generally kicking around ideas about how we might like to change the world. Again.
Previously we’ve mentioned our open access poster repository, and Cathryn has been talking about some of the more interesting and unusual posters that have appeared on the site. There was quite a bit of interest on twitter in the last couple of days, too. But naturally we’re not content to leave it like that: we’re also thinking about including raw data and figures, as part of the Posters repository and maybe beyond that. You already know what I think about supplemental data and whatnot in journals.
Mark Hahnel is a PhD student studying stem cell mobilization at Imperial College London, and founder of the Science 3.0 network. He’s built a new figure-sharing site, FigShare, which came out in beta today, and I invited him to write a guest post telling you all about it. Here’s Mark:
Scientific publishing as it stands is an inefficient way to do science on a global scale. A lot of time and money is wasted by groups around the world duplicating research that has already been carried out.
FigShare allows you to share all of your data, negative results and unpublished figures. In doing this, other researchers will not duplicate the work, but instead may publish with your previously wasted figures, or even offer collaboration opportunities and feedback on preprint figures.
Upload all their negative data.
Upload all their preliminary data.
…upload all data!
What percentage of the figures that went into your undergrad, masters or doctorate thesis were ever published? The ones that you didnt publish were probably good basic science, or figures that didn’t tell a complete story. So put them on FigShare (they’re already formatted): it takes just second to upload data.
Data can currently be uploaded as figures or as a dataset. The figures can be linked to datasets and vice-versa. Figures may include images, graphs, photographs and many other formats. Each figure or dataset will automatically have a reference generated, including a persistent identifier. This means that you can reference your individual figures in papers or presentations—and so can your readers, who will also be able to link the figures up with the relevant experiment.
Most researchers want their data to reach as wide an audience as possible; that’s why we present at conferences. (The other reason for going to conferences is to network and to find inspiration.) FigShare acts as an ongoing conference. Like F1000 Posters, the value for science is in the content: if you like the idea and think that the efficiency of science needs to be improved, then upload your data.
Some important characteristics of the FigShare database:
|
All data are easily SEARCHABLE You can browse the figures, either by ‘category’ in the sidebar or by each of the variables on the right. |
 |
 |
 |
 |
All data are CITABLE (with persistent identifiers):
All data are TAGGABLE (and all tags are linked, so all figures are linked):
FigShare also gives you the ability to easily share your figures and datasets via a host of social media platforms through ‘share buttons’ on every page. This is a new way of bringing scientific research online and to a new audience.
FigShare is a permanent database of your research. To guarantee this, FigShare is supported by the Systems Institute. Systems Institute is a not for profit organization that provides ongoing support for the hosting of FigShare as it expands. This also allows FigShare to make backups of all your data each and every day.
Through ongoing discussions between FigShare, F1000 posters and journal publishers, we now know that the majority of publishers are open to the idea of preprints. It is also now known that open access improves citation rates and the impact of research.
The beta release of FigShare is a fully functional data sharing platform. We will continue to develop the features of the site to bring you even more tools to help get your data to a global audience. If you have any suggestions or feature requests please get in touch, at mark@figshare.com, via @figshare on twitter, or through the FigShare facebook page. You can also leave a comment below.
|
|
previous post
Interesting concept. We all have a ton of data that has never been published. Much of it represents an isolated study or series of studies that does not sufficiently describe a ‘story’ necessary for publication in traditional journals . So, these ‘observations’ remain cloistered within a laboratory until someone, years later, may or may not pick it up and run with it. Why not let someone else run with it? Worse, these same preliminary studies are repeated over and over again elsewhere.
My specific comments surround the execution of this concept.
-What is minimally acceptable as a ‘figure’ in order for it to be posted? Does one allow anything to be posted? Or are there certain criteria that must be met (minimum number of independent replicates with reports of statistical significance)? Does one allow multiple linked figures that now look more like a paper than a isolated figure? At what point does ‘data sharing’ become ‘non-reviewed publication’?
-If the site is even moderately successful, data searching may be overwhelming. But the counter-argument is that it is better to spend a day doing an exhaustive search than a year developing the material and conducting the study that someone else already knows the answer to.
-The stability of the data storage is a key, particularly if this morphs into non-reviewed publication. The beauty of libraries (physical or electronic) is that the work becomes distributed in multiple locations. Even if the site of the data repository is rock solid, it still would be smart for this initiative to ensure that the data is distributed beyond a single site succeptible to loss.
Hi Fred,
I’d like to try to answer your questions on behalf of http://FigShare.com
A figure can be any level of scientific data, be it n=1, a preprint figure or a previously published figure of yours. The value is added by the tagging system and ability to share through various platforms. This gives other researcher the opportunity to find data relevant to theirs and contact the author. Non-reviewed publication would require some form data analysis, FigShare just provides the raw facts, interpretation is dependent on the viewer.
By forcing users to tag their research uploads, the data is easily searchable, the more tags you add , the more specific the search can be. As it stands, a figure published within a paper is almost un-google-able. By providing a service that includes figure titles in searches as well as categories and tags, the figures are now easier to find than within established publishing models.
Back ups are created daily and stored both internally in Europe and duplicated on servers in the US. Access to all figures is permanent due to persistent identifiers from handle.net. Thanks for the feedback, please get in touch if you have any more questions or ideas.
Cheers.