nfais webinar scholarly collaboration networks

17
Scholarly Collaboration Networks William Gunn Director of Scholarly Communications, Mendeley http://orcid.org/ 0000-0002-3555-2054

Upload: william-gunn

Post on 08-Feb-2017

206 views

Category:

Science


2 download

TRANSCRIPT

Altmetrics trends

Scholarly Collaboration NetworksWilliam GunnDirector of Scholarly Communications, Mendeleyhttp://orcid.org/0000-0002-3555-2054

Hi, Im William Gunn, Director of Scholarly Communications for Mendeley. In the next 30 minutes, I would like to briefly discuss how Mendeley fits into the scholarly collaboration ecosystem and what were doing to support researchers as their networks move online.1

Mendeley: Used more for papers than people.5+ million usersTwice as many users use Mendeley dailyAverage account age is twice as oldhttp://www.nature.com/news/online-collaboration-scientists-and-the-social-network-1.15711

There was a survey done by Nature News last year which looked at social networks used by scholars. They looked at Twitter, Facebook, LinkedIn, Mendeley, Researchgate, and Academia.edu. It was unclear in some cases whether researchers were responding based on their scholarly use or general use of the platforms, and they tried to focus on the social use only, which left out a major part of the Mendeley demographic, but the data are interesting in how the illuminate the implied needs of researchers. In this survey, they found that academics primarily use Mendeley as a tool to share and discover research. Other interesting bits of data are that twice as many researchers use Mendeley daily as the other scholarly sites that are more along the lines of Facebook for research. In addition, the average age of a Mendeley account is 2 or more years old, whereas with the other scholarly services, the age is closer to a year, despite the services all being around for around the same length of time. Mendeley launched in 2008. This is likely due to what we call churn, where people try a service for a little while but dont stick with it if they dont find it useful.2

Twitter: All about the discussionSocial media activity around scholarly content growing 5-10% per monthAdie, E. & Roe, W. (2013) Altmetric: Enriching Scholarly Content with Article-level Discussion and Metrics Learned Publishing 26(1) 11-17http://www.nature.com/news/online-collaboration-scientists-and-the-social-network-1.15711

On Twitter, its clearly all about the discussion, though theres some discovery of novel research or research peers as well. Social media discussion of scholarly content has been growing steadily at a pretty good clip, though the discussion primarily remains focused on breaking news such as natural disasters, health or medical research of general interest, for example about stress or dieting, and a fair amount of jokey stuff about cute animals or funny research topics. Such is the nature of the internet. If you want to be big on Twitter, study the effects of marijuana on cute kittens.3

LinkedIn (and Academia/RG): ProfilesMostly passive use with a high ratio of monthly use to daily usehttp://www.nature.com/news/online-collaboration-scientists-and-the-social-network-1.15711

We see a different, almost opposite, usage profile for sites that are more about the profile or online CV. Researchers do use these sites, but mostly passively. They create a profile and then leave it alone, checking in maybe once a month or so.4

Facebook: Not used professionallyhttp://www.nature.com/news/online-collaboration-scientists-and-the-social-network-1.15711

Curiously, Facebook is mostly unused for scholarly communication, with most researchers saying that while they check it daily, they dont do so for professional reasons.5

Main use cases for scholarly networks Sharing researchDiscussing papersHosting an online profileDiscovering researchSearch, discover, read, reviewSynthesize/AnalyzeExperimentRecruit/evaluate researchersSecure FundingManage facilities

Publish anddisseminateManage DataPromoteCommer-cializeCollaborate & networkEstablishpartnerships ?Develop StrategyHave impact!

OK, so what does all that tell us about how researchers are working together online? For us at Mendeley, it looks something like this. There are many, many activities researchers do as part of their professional life beyond do experiments and write papers. There are a range of activities they do, loosely under the headings of preparing, doing, and communicating, such as writing grants, recruiting/evaluating other researchers, setting up collaborations, searching, reading, & reviewing papers, analyzing data, as well as explicitly networking (online and offline). From the survey data, the services that most researchers are availing themselves of are focused on sharing research, discussing research, hosting an online profile, and discovering research. Ill now go show what Mendeley is doing to support these needs.6

Mendeley groups make it easy to share papers, write manuscripts together, and discover related research.

Groups can be interacted with via other applications via the groups API.

Just as researchers use Mendeley to organize papers for themselves, they use it to share research with others. In fact, weve implemented sharing such that its essentially a social folder. To your folder of papers, you can invite colleagues discuss papers, and write papers together - working from the same set of references. You can also find groups that researchers have made public, so its worth taking a moment to look for a group on your topic rather than creating your own. Once youre in a group, you can see groups that are related to it, adding further serendipitous discovery opportunities.7

In addition to basic info about interests and publications, profiles show usage of those publications.This information is also available via API, so it can be leveraged in other applications.

We do have profiles at Mendeley, but in addition to basic info such as research interest, publications, and affiliations, we also show useful metrics about your publications. One key feature of a Mendeley profile is that once youve created it here, you can use it across Elsevier services, such as peer reviewing and manuscript submission, and third-party services can also leverage this information so a researcher doesnt have to update his information separately at each site, instead those sites can be updated automatically with any changes here.8

Leveraging the networks of Mendeley users, but also the implicit networks of co-readership, co-authorship, co-citation allow great recommendations to be made.

Currently showing content-based, popular in your field, and trending in your field.

This information is also available via API, so it can be leveraged in other applications.

Leveraging the networks of Mendeley users, but also the implicit networks of co-readership, co-authorship, co-citation allow great recommendations to be made. Many times, a related paper wont use the same language or keywords as the paper its related to, making it hard to find the connection. By looking at papers that co-occur in research libraries, we can find these hidden relationships. We use this to show a carousel of recommendations for your library or your field, as well as on-demand recommendations for a set of papers. If youll recall, the main thing researchers said they liked Mendeley for, among the limited set of things the survey asked, was the recommendations.9

A new service from Mendeley is focusing on research data. Heres what a dataset summary page looks like at Mendeley Data. Note that each dataset has its own DOI, can be cited independently of any paper which may or may not exist, and its also preserved in DANS in case anything were to happed to cause Mendeley Data to be unavailable. In this example, there is a published paper which references this dataset, show in the blue box above.10

Heres what it looks like from the page in ScienceDirect for the referencing article. You can see that the article is linked to the data on Mendeley Data, which is SO MUCH better than sticking it in supplementary data where its hard to get to and much of the data-specific metadata is lost. For example, the code used to generate the future or summary tables from the raw data can be uploaded to Mendeley Data (or linked to from there), which is a MAJOR step forward for reproducibility and provenance in research data management.11

My preference would be for this system to eventually replace supplementary data entirely. With a flexible system for managing and storing data, we can show interactive visualizations for data right on the article page, which means the researchers dont even have to download the data to do basic manipulations they can play with it right in the page.12

ChallengesAdoption building critical massFunding having support to dream and buildSupport helping each other

There are a number of challenges that scholarly communication networks face as adoption grows. The first is making sure that as adoption grows, these services are actually solving problems for researchers and not creating more work or distraction. Another challenge is incentivizing the creation of new products and services for researchers through a diverse range of funding sources that attracts the best talent, not just the talent thats able or willing to sacrifice the big money for idealistic reasons. Finally, support is also needed from the institutions which host the researchers and among the various services themselves, such as my fellow companies in this webinar with me.13

AdoptionFulfilling a useful purpose for researchersAdoption continues to growHow do we avoid creating extra work?Utilize existing infrastructure, such as ORCID, DOIs, CSL

With respect to adoption, we really have to focus on the needs of the researcher and avoid the temptation to build something just because it worked somewhere else for a different population. Twitter is doing just fine, we dont need Twitter for Research, but given the reluctance of researchers to use Facebook for professional means and the failure of LinkedIn to serve this space, perhaps Facebook for Research isnt such a bad idea after all. Important as we work on these things, though, is to leverage common infrastructure where possible. No one should be creating their own way of rendering citations or creating their own researcher IDs. We have infrastructure for this and we should all use it to reduce not only the burden on researchers, but the burden of development of new services, too.14

FundingPublishers shouldnt remain the main source of fundingVenture capital needs to look at this spaceAcademics need to become more comfortable with for-profit business modelssustainabilitypreservation

It is going to be a problem is the main incentive for a bright young team of developers to build an application is to be acquired by an existing publisher. Disruptive models tend to come from without the industry and if startups are positioning themselves for acquisition by a publisher, its going to hurt innovation. As a corollary to this, academics need to become more comfortable with for-profit business models. They largely are already with Facebook and Twitter, but if advertising isnt acceptable in the scholarly space, we need other models, and certainly paid services are better positioned to guarantee important things like user privacy, data stewardship, and service sustainability.15

SupportPublishers should make it easy for third-party platformsThe STM Association has released a set of scholarly sharing principles which affirm the right to share on third-party platforms. http://www.stm-assoc.org/stm-consultations/scn-consultation-2015/Institutions need to understand how to support researchersLibrarian Certification Programhttps://www.elsevier.com/solutions/mendeley/support/mendeley-certification-programAll stakeholders need to agree on which bits of the scholarly communications apparatus must be community-owned, funded, and managed and which should be commercial. Crossref and SHARE are model organizations in this regard, focusing on infrastructure, not end-user services.

Finally, support is needed in various ways from all the stakeholders here. Publishers need to do their part to work with third-party platforms, and Im happy to say that Mendeley has led the effort from within that resulted in the STM Association affirming the right of sharing. All stakeholders need to agree on which bits of the scholarly communications apparatus should be community owned and which should be commercial. I would nominate Crossref and SHARE as model organizations in this space (Crossref being significantly more long-lived) in that they focus on infrastructure the plumbing and roads of scholarly communications not end-user services. This will accelerate innovations and better serve the researcher in the end.16

www.mendeley.com

[email protected]@mrgunn

I hope this has been interesting and informative. Please feel free to get in touch with questions via email or tweet me @mrgunn. Thanks for your attention!