Saturday, June 07, 2008

Maps of Science

Interesting website visualizing scienctific research fields based on a co-citiation analysis of journals.

Unfortunately, one can't zoom into the fields, that would be really cool.

I came across the site via this paper
    Mapping the backbone of science
    By Kevin Boyackm Richard Klavans and Katy Boerner
    Scientometrics, Vol. 64, No. 3 (2005) 351.374

which is quite visionary in its aims. Here is a quote from their conclusions:

"The disciplinary map presented here is designed to support decision-making, e.g., the allocation of resources among/between disciplines. However, it also promotes the understanding and teaching of the general structure of science. Although it is a static map, and thus does not reveal how disciplines are born, evolve, or die, it is the broadest static map of science published to date, and thus constitutes another step forward in the study of the structure and evolution of science by scientific means.

Ultimately, maps of science could be based on a much broader set of data (such as scholarly journals, proceedings, patents, grants, and funding opportunities). Alternative units of analysis (clusters of journals, papers, authors, funding sources and/or text) could be generated to address different user needs. Instead of being static, dynamic maps could be generated that show high activity, scientific frontiers, and merging/splitting of scientific areas.

We believe that these global maps of science will enable researchers and practitioners to search for and benefit from results and expertise across scientific boundaries, counterbalancing the increasing fragmentation of science and the resulting duplication of work. These maps of science could also serve as a common data reference system for scholars from all disciplines - analogous to how geologists use the earth itself to index and retrieve data, documents, and expertise, or to how astronomers use astronomical coordinates. If such a reference system were to exist, all researchers could have a bird's eye view of the landscape of science, and could use this landscape to navigate to areas of interest, to communicate results, and to announce discoveries.

This global view - as opposed to doing keyword based searches on the Web or in digital libraries with very little information about the coverage of the queried database or the quality of the result - would give many more people access to scientific results. This, in turn, would lead to more informed citizens and a faster spread of results and practices benefiting all of humanity."


  1. Hi Bee,

    A truly interesting idea and related web site and for sure something that requires careful assessment and contemplation before comment. With the quick peak I had of it I can point to only one complaint which is the same I have for light beer and that it takes far too long to get loaded:-) I can see for me to utilize and explore this properly I require a faster computer. Mine now is two years old and with this you have convinced me at least in certain instances it is inadequate. This post could therefore prove expensive.



  2. Hi Phil,

    I am a dedicated hater of flash, so my apologies, I could have warned you that site takes forever to load for no good reason as far as I can tell. I like it for the idea, not for the presentation. Not a reason to get a new computer. To make a speculation, I think that many overloaded sites are likely to slim down on their loading requirements, as it is to the annoyance of the visitor, a fact that must sink in with webdesigners at some point.

    In this regard, you might find this article interesting that I read yesterday on BBC News

    Web users 'getting more ruthless'



  3. Web users aren't so much more ruthless as tired of people's bullshit.

    This is a much better science map.

  4. Hi Eric,

    thanks - the map you link to refers to the same oirignal data, it seems.

    It's a pity that there is no "reasonable", usable version of the map available, it's content is really interesting!

    Best, Stefan

  5. Hi Eric, Stefan,

    Yes, it is the same map. I came to the above mentioned website via this site where you can download a high res. version (watch out, it's some MB). Best,


  6. Blaaaaah. I had the high res version where you can read everything on my drive, but its 5.3 mb. So I googled the file name, and used the first thing that seemed reasonable.

    I like that version better.

  7. I wouldn't expect such a map to be complete, but some things seem rather wrong (surprising). For example, there is a huge literature in bioinformatics/computational biology (my field) that should intersect biotech with CS/math - yet there doesn't seem to be a single link.

  8. Hi Anonymous,

    Yes, there are also some other aspects I find unintuitive. There is two things one should take into account however. One is, the links shown have to cross a certain threshold, so there might be a link but if its not pronounced enough it isn't shown. The other is that one has to keep in mind what it is the map shows. It's not keywords or areas, it is journals and co-citations. Two journals can be closely related in topics, but if the topics very neatly classify the links won't be very pronounced. E.g. when it comes to mathematical modelling it could be a field more or less develops its own procedures that are very specialized in application. I could be that the use of research on general cs topics is not actually very high, despite the fact that it is the underlying input. I don't know anything about computational biology, so this is just a speculation. It is certainly the case if I look at numerical models that have been developed and used in theoretical physics. You'd say that's software development, computer science, and so on, but it's only a very specific audience that's addressed with this research. Best,


  9. Hi Bee,

    “I don't know anything about computational biology”

    This will not do. Biology is obviously far too difficult to be left to the biologists!

    Regards, Dany.

  10. Alan Turing -- who, IMO, is one of the finest, yet most underrated, thinkers of the 20th century -- offers an easy explanation as to why it's not gonna be easy going from bits to qubits:

    "There is a remarkably close parallel between the problems of the physicist and those of the cryptographer. The system on which a message is enciphered corresponds to the laws of the universe, the intercepted messages to the evidence available, the keys for a day or a message to important constants which have to be determined. The correspondence is very close, but the subject matter of cryptography is very easily dealt with by discrete machinary, physics not so easily."


COMMENTS ON THIS BLOG ARE PERMANENTLY CLOSED. You can join the discussion on Patreon.

Note: Only a member of this blog may post a comment.