Peter Binfield: Article level metric from the... - SciBarCamp

What is the problem: how do you access the worth of impact of journal articles? - Martin Fenner

What is the granularity to measure impact: journal, research institution, individual researcher? - Martin Fenner

important decisions made on the basis of "impact" which covers many things, interest, usability etc - Duncan Hull

So much to read need metrics to winnow things to read a lifetime. - Bosco Ho from iPhone

We talk about "metrics", but many signs of impact are hard to measure - Duncan Hull

Usage-based measures are still pretty rare - Martin Fenner

Current metrics include impact factor, citation based measures (h-index) - Duncan Hull

Most measures are journal level - Duncan Hull

Dowload stats - Duncan Hull

There is no substitute for reading the literature yourself and forming your own opinion - Duncan Hull

Most people filter based on the impact factor of the journal. - Martin Fenner

All metrics are terrible, some metrics are more terrible than others - Duncan Hull

Potential metrics: citations, usage, discussions, social bookmarks, etc. - Martin Fenner

Martin made a great point: Why should we even want to assign a number to the quality of a paper? - Chris Patil

Web 2.0 is the solution. - Bosco Ho from iPhone

People are starting to use these, but not very much, so PLoS ONE wants to start the ball rolling. - Chris Patil

PLoS provides article-level metrics for all their articles, e.g. citations, usage, blog posts, etc. - Martin Fenner

there's some good arguments against assigning numbers to students in school, also. I think stats are a good thing, as long as they aren't the only thing. Remember, Mendeley http://www.mendeley.com/stats also is collecting usage stats, in which PLoS figures quite well. - Mr. Gunn

article level metrics fall into six areas 1. citation metrics 2. usage metrics 3. expert rankings (F1000) 4. Conversations (blogs, media coverage, comments) 5. social bookmarking (citeulike etc) 6. Other cools stuff (geotagging of authors etc) - Duncan Hull

Number of blog posts about an article isn't necessarily a measure of quality - a flawed paper might attract more commentary and blogging than a flawless one. - Chris Patil

Article-level metrics as post-publication peer review. - Martin Fenner

pre-publications and post-publication metrics - Duncan Hull

Need open APIs to access the data (not proprietary) - Duncan Hull

@patil a flawed paper could potentially be more interesting. - Bosco Ho from iPhone

ResearchBlogging.org are building an Open API at the request of PLoS one - Duncan Hull

usage numbers going to be available soon at plos - Pedro Beltrao

Example http://www.plosone.org/article... - Duncan Hull

Use open API data of other web sites to aggregrate stats. - Bosco Ho from iPhone

click on "related content" http://www.plosone.org/article... - Duncan Hull

citeulike more heavily used than connotea according to plos one - Duncan Hull

Citeulike >> connotea - Bosco Ho from iPhone

very few comment or add ratings - Duncan Hull

Let people comment in whatever space they feel comfortable (friendfeed, blogs, disqus, citeulike, whatever etc) and then aggregate that content. e.g. www.postgenomic.com requires use of rel="rev" tag in html - Duncan Hull

Future work, integration with mendeley, zotero, papers, researchblogging.org add usage data (coming in August) - Duncan Hull

Plosone is big enough to twist the arms of big sites to provide an open API. The possibilities are endless. - Bosco Ho from iPhone

friendfeed and twitter discussions difficult to aggregate, rarely mention DOI or authors or title of paper :-) - Duncan Hull

Don't expect conversations on http://www.plosone.org anymore, this is crucial - Duncan Hull

BUZZWoRD alert: crowdsourcing. - Bosco Ho from iPhone

question from Alex Tolley https://twitter.com/atolley How are you going to do it? - Duncan Hull

CrossRef is working on the heuristics of a citation, e.g. good or bad paper. - Martin Fenner

Usage statistics will show trends over time. - Martin Fenner

Otherwise only bmc provides download stats for authors limited data for public. - Bosco Ho from iPhone

Usage statistics is extremely time-consuming, largest EC2 instance is working for 9 days already on PLoS One usage data. - Martin Fenner

What is a "big number" of downloads? Difficult to know because much of this data is currently hidden - Duncan Hull

Tufte fanatic is critiquing the aesthetics of a beta release graph. - Bosco Ho from iPhone

Still critiquing. - Bosco Ho from iPhone

Currently no open API for PLoS One usage statistics. - Martin Fenner

This man really cares about aesthetics - his apartment must look fabulous. - Bosco Ho from iPhone

Future developments at PLoS: provide open data sets, hope other publishers will follow, hope standards will evolve - Martin Fenner

Article statistics on plos one throws down the gauntlet against elsevier and friends - Bosco Ho from iPhone