Multiple Metrics for “Visions of a Semantic Molecular Feature”

After posting the access metrics for VoaSMF , Egon Willighagen suggested that we also use some of the new alternative metrics (“altmetrics”). There’s more than one such effort and they are to be welcomed (citation metrics by themselves are inefficient, imprecise and suffer huge timelag). So people such as Egon, Heather Piwowar, Cameron Neylon are creating immediate metrics – literally hour-by hour. Here we show “Total Impact” (

Before we show the figures let me commend this and similar efforts because:

  • They are not tied to a commercial business. (Journal Impact factors have long been tainted with the suspicion that they are manipulated between aggregators and publishers. Not for the benefit of science, but in the near-meaningless struggle between journals for branding)
  • They are immediate. Within hours of publication you can get alternative metrics.
  • They are objective and democratic. Anyone can build tools to carry them out. It gives real scientists a say in how science is measured. I expect universities to give them the cold-shoulder for some time as they challenge the current trivial system and they are free.
  • They are multivariate. A whole host of different metrics can be used. You can make your own up.

The altmetrics software was originally hacked at a workshop that Cameron ran, I think. Anyway it is typical of the quality and speed that can be achieved by people working together with a common vision and shared tools. Indeed (I hope) this shows the challenge to the lumbering publication systems that publishers build and force us to use. We are starting to liberate our expressiveness.

So here are our 15 articles and I comment later:


report for Visions of a Semantic Molecular Future

download data
run update updated 15 Nov, 2011 created 15 Nov, 2011 15 artifacts



Dataset [there is a bug, these should be labelled “article”]

More detail on available metrics. Missing some artifacts or metrics? See current limitations.

status log

an project.
source code on github

There are at least 6 metrics:

  • Tweets
  • Blogs (a limited selection)
  • Bookmarks (Cite-U-Like)
  • Wikipedia entries
  • Facebook shares and likes
  • Mentions

Most are sparsely populated – the exception being JohnW’s tweets. These are real – the twittersphere resounded on day with a massive list of tweets about John’s article. There are some technical issues – some metrics currently require DOIs, etc.

People may argue that these metrics can be gamed. (Of course John pulled people off the streets of SF to tweet his article.) Seriously I think the accesses are reasonably accurate and haven’t been gamed AFAIK. The altmetrics don’t visit enough blogs, I think but that will come.

The point is that rather than waiting 3 years to find out if anyone has read our articles we get a current picture. It doesn’t surprise me that there are hundreds of accesses for each tweet or blog – we’ve had 300,000 downloads of Chem4Word and hardly a squeak. And most cheminformaticians don’t tweet or blog or show themselves in the glare of Open social networks.

Thanks to the altmetricians for their efforts.


This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *