Altmetrics will be taken personally at PLoS

altmetrics12
ACM Web Science Conference 2012 Workshop
Evanston, IL, 21 June 2012

Martin Fenner
mfenner@plos.org
Public Library of Science, USA

Introduction

Altmetrics have traditionally looked at metrics relating to individual scholarly works, typically journal articles, but increasingly also including datasets and other research outputs. This has become feasible only fairly recently thanks to advances in technology and is an important step forward from traditionally used aggregated metrics such as the Journal Impact Factor. The Public Library of Science (PLoS) has been providing Article Level Metrics (ALM) for all their journal articles since 2009, making them arguably the largest provider of altmetrics for scholarly papers. PLoS is providing data on article usage, blog and media coverage and social networks in addition to traditional citations, and all this information is freely available via website or API.

Central goals of the altmetrics movement are to provide a) a more nuanced picture of impact, beyond citations and b) better filters for finding scholarly content. Both of these goals can only be achieved if the metrics collected for scholarly works are connected to both creators and users of these works. In this paper I propose that personalizing altmetrics is essential for every altmetrics project, and I will discuss the steps PLoS is undertaking to personalize the PLoS ALM project.

Personalizing ALM at PLoS

The PLoS ALM project is tracking a variety of metrics about PLoS articles that contain identifying information:

Source User role Category Comment

CrossRef

article author

Citation

requires disambiguation

PubMed Central

article author

Citation

requires disambiguation

ResearchBlogging

blog author

Citation

PLoS Comments

comment author

Comment

Twitter

user

Comment

in testing

Facebook

user

Comment

Bookmark

comments, likes and shares are aggregated

CiteULike User

user

Bookmark

CiteULike Group

group member

Bookmark

Mendeley Group

group member

Bookmark

Most of this identifying information is currently hidden and unconnected. In order to personalize the PLoS ALM content, we are planning the following improvements:

  1. Add user information to all sources, where possible
  2. Disambiguate user information
  3. Add user accounts
  4. Aggregate user profiles from different sources
  5. Use user profile information
  6. Build reporting and recommendation tools

User information can’t be provided for all sources, primarily because of privacy concerns, licensing restrictions, and because that information would sometimes be very difficult to obtain, e.g. usage stats or all authors of a Wikipedia article. Some sources can still provide relevant user-centered information. Mendeley provides anonymized reader information for discipline, academic status and country and usage stats can be filtered by IP adress to provide location information.

Some sources provide unique person identifiers, for all the other sources this information needs to be disambiguated. This is a complicated issue in particular for authors of scholarly papers. The Open Researcher & Contributor ID (ORCID) initiative aims to solve this name ambiguity problem for scholarly authors. The service will launch in the Fall, at which time PLoS will start to ask authors submitting manuscripts to provide their ORCID identifier. Authors of science blogs often link their blogs to their Twitter account, and they could also use ORCID identifiers.

Personal user accounts can provide more relevant filters for scholarly content. This would enable features such as tracking the metrics of bookmarked papers and saved searches, and of course of all PLoS papers authored by the user.

Many social media sites aggregate user profiles from different services, but they usually don’t integrate scholarly identifiers used in some of the metrics collected by PLoS. It is unclear at this stage whether the ORCID service will also link to social media profiles on Twitter, Facebook, etc. The best places for aggregation of the various user accounts are still unclear, and could also include journal websites, including PLoS.

User profiles cannot only be linked to the metrics of the scholarly works they authored and to profiles in other services, but they contain valuable additional information. The most important information is the connections to other users: as coauthors of scholarly papers, friends or followers of other users, or because they belong to the same user group in social networking sites.

Not all aggregation of identifying information has to take place in the ALM application, but once this identifying information is better integrated, we can use this highly connected network of users and articles with their metrics to build better impact reporting tools and recommendation systems. Some of these approaches are already in use by services such as CiteULike, Biomed Experts, Microsoft Research and VIVO.

As an Open Access publisher with a rapidly growing corpus of articles PLoS has the unique opportunity to not only develop these personalized metrics for its journals, but to also make this information available to other researchers to answer important information science research questions.

Creative Commons License

This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.

Post a Comment

Your email is never shared. Required fields are marked *

*
*