Related name pages

I've done other posts about related names in Identities records (e.g. Identity cycles and Forgotten people).  Related names are typical of the sort of thing that can naturally link two different ways.  Right now in Identities we make the main link to the Identity page for that name and add a [+] after it that does a search in WorldCat.org for both names in an attempt to show people how the names are related.  Here's an example from the Lewis Carroll page:

Related Names

• Tenniel, John Sir 1820-1914 Illustrator [+]

The Tenniel link goes to a page about John Tenniel in general, but the [+] link searches WorldCat for John Tenniel and Lewis Carroll.

The search can be useful, but I've been wondering whether pages devoted to a pair of names would be useful.  I also wondered how many there would be.

I don't have the pages yet, but I do have some statistics.  The 21 million personal name pages in WorldCat Identities have more than 47 million related names on them, but only 36 million of these pairs are unique (since many Identity pairs refer to each other).  Since the pair pages for people with few citations don't seem very interesting I eliminated pages with less than five citations from consideration.  I also eliminated relations where they have only a single citation in common, since they didn't seem interesting either.  After all that there are a bit more than four million pair-pages.

I'm not sure exactly what such a relationship page would look like.  In some ways this is data there has never been an easy way to see before, but I think they would be interesting.

--Th

Note: Dedicated users of Identities will have noticed that the system has been either unavailable or unreliable the last few days.  Our server was 'compromised' and it has taken us a bit to recover.  We should be back to normal soon.

Comments

[David]

Interesting idea: since no one else has commented, here are belated random ideas.
One straightforward picture would be a paired timeline, e.g. one author at the top and the other timeline underneath, reflected so that their publications extended downwards. Titles with which they were jointly associated could be coded so that they were visible within each of these timelines (i.e. counting them twice, to make a symmetric pattern above and below the line). Here the interface/overlap between the two individuals is a cluster around a horizontal line, though info of variable length might more naturally be comparatively arranged vertically.
Another idea: a column of tags relating to X only , a column of tags relating to Y only, and in between (either as a column or more cunningly spaced according to the extent to which they were shared) tags which applied to both.
You can clearly identify cases of 'X wrote about Y', and also have the MARC relator codes. I wonder how pairs of these would cluster in your dataset. I've often thought it would be wonderful if there were some way to pick out authors who wrote books attacking other authors, or who were paired as controversial disputants, but information about that sort of polarity of association unfortunately isn't there in MARC. You're right that this is data it hasn't been easy to isolate as a focus of attention before, and hence we don't have well-developed ways of seeing or labelling the sorts of relations there are.