What if journal current contents were tag clouds?

Would it tell us anything useful? Does it tell us anything useful about the current state of journal RSS feeds, or suggest areas for improvement?

Let’s find out. It’s as simple as locating the feed URL from your journal subscriptions, pasting it into Wordle and eyeballing the results. Here’s what came back for the 25 journals in my current subscription list.

Biochemistry
Not especially successful. Items in the Biochemistry feed consist largely of dates/times, DOIs and author names, with no information regarding article content.
Bioinformatics
Rather better. It’s obvious at a glance that protein kinases and phosphorylated substrates are topical in this issue.
BMC Bioinformatics
Also very good. We would guess straight away that analysis of proteomic data is a hot topic in this issue.
BMC Genomics
Another win for a BMC journal. It’s clear that gene expression is big in this issue.
Briefings in Bioinformatics
This graphic gives us a flavour of what the journal is all about: more methods and algorithms than insights into biology. Nothing wrong with that, just saying.
Cell
Networks stands out, but then we see a lot of words with similar frequency (most likely, one instance). Not especially informative in this case.
Genome Biology
I’d guess that this is a special issue on genes controlling circadian rhythm in Drosophila. An excellent summary graphic.
Genome Research
Not too surprisingly genes, genomes and genomics feature strongly in this summary. We also see an emphasis on mammals and an intriguing Mexican connection.
Genome Research in advance
Articles in advance from the previous journal. Once again, it looks like we have a special focus on Drosophila.
Journal of Biological Chemistry
An excellent summary from JBC. We’d expect to find a number of current articles on membrane channels, judging by this graphic.
Journal of Biochemistry
The second of the biochemistry journals also scores well. Lectins, the extracellular matrix and fatty acid metabolism are the ones to watch this month.
Journal of Proteome Research
Oh dear – we’re back to dates, DOIs and authors. The feed from this journal contains pretty summary graphics for each article, but tells you little about content.
Nature
Looking at this, you might get the impression that Nature rather enjoy talking about themselves ;)
This one is a little weird in that the feed clearly contains abstracts, but whatever I gave to Wordle doesn’t extract the words as expected. See Alf’s comment for something that did work well.
Nature Biotechnology
Same problem for another journal from the Nature stable.
Nucleic Acids Research
Gene expression and neuroscience feature in the latest edition of NAR.
Nucleic Acids Research in advance
Advanced access articles from the previous journal are also summarised nicely. Reassuring to see that nucleic acids still feature prominently, given the title of the journal.
PLoS Computational Biology
It’s very obvious from this summary that the journal is all about models. There seems to be a focus on malaria in the current issue too.
PLoS ONE
A strong showing from PLoS ONE and as you might expect, a wide range of topics. “Oryzae” is prominent in this summary – rice, Aspergillus, other?.
Proteins: Structure, Function and Bioinformatics
Reassuringly, a range of keywords related to protein structure appears here. A PDB code, 1TW7, is also prominent.
Proceedings of the Royal Society B
A decent-enough summary of current content. Fruit anyone?
arXiv Quantitative Biology
I like to keep an eye on the biological section of arXiv. As you’d expect if you’ve ever browsed the contents, an eclectic mix of concepts are captured in this summary.
Science
Disappointing. Plenty of keywords but none that really jump out: largely because many of them refer to the section titles used by the journal. By now, you should be realising what makes for a good RSS summary. Hint: it’s the abstracts.
Source Code for Biology and Medicine
The neglected, lower tier end of the bioinformatics journal spectrum, but generates an excellent summary graphic. Clear that image analysis is the hot topic in this issue.
Structure
Not a huge amount of variation in keyword frequency here, but it at least gives us some impression of the contents. S-layers are the only stand-out.
Trends in Biochemical Sciences
There’s a joker in every pack, isn’t there.
Frankly, with a URL to the RSS feed that looks like this (please don’t click – just hover over it and marvel), it’s best not to expect much from Science Direct.

In conclusion

  • The best RSS summaries from journals contain abstracts
  • Cluttering up your feed with section headers (often longer than the article titles) reduces readability and information content; I’m looking at you JBC and Science
  • Pretty pictures in feeds contribute nothing

14 thoughts on “What if journal current contents were tag clouds?

  1. Lars Juhl Jensen

    Nice idea, just a shame that so many journals have RSS feeds that do not include abstracts. Have you considered making your own RSS feeds for each journal with PubMed and analyzing those instead?

  2. Pingback: nsaunders does a tag cloud on journal feeds « My Weblog on Bioinformatics, Consed, Phrap, Genome science.

  3. aboulia

    nice one!
    going toblog about this..
    erm curiously.. why isn’t there a tag cloud for your blog? ;p

  4. Scott Kerr

    Nicely done, I find it very irritating that many journals don’t send out good info in their RSS feeds. Whats the point of sending it out without any info? It’s 2008, they should have this figured out by now.

  5. nsaunders Post author

    Back to the Nature feed – various sources suggest that Wordle uses only the first 20 items. If so, that wouldn’t get us to the papers in Nature and may explain the problem.

  6. Pingback: What’s on the web? (29 August 2008) « ScienceRoll

  7. Judson Dunham

    Hello Neil, Judson Dunham from ScienceDirect here. First of all, great idea!

    You’re absolutely right about our RSS feeds–they should contain abstract information. Unfortunately a very pernicious bug has prevented most of the TOC feeds from displaying abstracts. This what the feed should look like: http://rss.sciencedirect.com/getMessage?registrationId=GIEAHLMAHKEIOIIEIIEIGQGHGMHGJJNAOAGIJOMEGS. (This is a search RSS feed that contains only Trends in Biochemical Sciences articles.)

    We’ve finally identified the problem and it’s planned to be fixed in mid-November with the next ScienceDirect release. At the same time we’ll be fixing the long URLs with a system of much simpler, stable URLs, and making the RSS feeds easier to find from within ScienceDirect using standard browser functionality. With the format and labeling we use they still won’t look good when plugged into Wordle, but hopefully they will be clear and useful for everyday subscribers.

  8. nsaunders Post author

    Hi Judson – I do appreciate your input. I’m sure publishers get tired of being bagged out by bloggers in dark corners of the internet! Looking forward to the improvements.

Comments are closed.