Is Scimago a good way to rank conference impact?

Question

I have seen that the Scimago Journal Rank make a rank of journals that cover different topics and classified them into quartiles. For what I know conferences also made their proceedings books, but I was dubious if they could be compared agains journals like Scimago does.

I say this because I believe that journals have most ot the time a higher impact than conferences, and compare them side to side (journals and conference proceedings) it is not such a good idea.

The question that I have is if Scimago could be a good way to rank a conference impact, and also if there is another computer science conference ranking that uses the same quartile calculus as Scimago?

Any help?

Thanks

Answer 1

Is Scimago a good way to rank conference impact?

No. It isn't.

At best, Scimago is a good way to obtain a modified "PageRank" of a publication in the graph of citations between Scopus-indexed publications in a three-year window, with each citation weighted by the similarity of the citing and cited publications, as measured by their common citation profiles.

Even if you accept that Scimago's abstruse formula is an accurate indicator of "impact"—which is debatable for numerous reasons—neither the raw citation data nor the precise definition of "cocitation profile" (on which the formula depends) are available to independently verify Scimago's rankings.

In particular, a few spot checks suggests that Scimago's coverage of major computer science conferences is spotty, and that the data it extracts from those conferences (even for relatively straightforward things like "number of citeable documents") is not particularly accurate.

Answer 2

This depends a lot on particular area. In some areas, proceedings of the best conferences are comparable with decent journals, while in others conferences are more social events. Computer Science in general is more a conference-oriented area, so proceedings usually have decent impact.

To check it in particular subarea, if you have access to Scopus (other services probably have similar features), try the Analyze Journals tool, to compare some conferences and journals in your field. In my case (Computer Vision), the top conferences are beaten only by the top journals.

For more info on different habits in different subareas I recommend: Wainer, Eckmann, Goldenstein, Rocha: How Productivity and Impact Differ Across Computer Science Subareas. Communications of the ACM, 2013.

edit: I just re-read your question and noticed your actual answer was about something else. There I can point you again to Scopus (which has both journals and top conferences) and their SJR/SNIP...