6
$\begingroup$

Often we combine related graphics to help us better understand a problem. We do this for poster talks in research (though we include text). Many info-graphics also do this as do this as do dashboards.

The problem is I don't know the proper term for Combining related graphics, including static graphs.

lattice, faceting, or trellis graphics (A subset of what Tufte called multiples) in which data is split into two or more grouping variables, are not what I'm talking about as each facet is the same type of graph as the others. What I mean is the act of combining related graphs usually without text as we do in a poster.

Here's an example in ggplot2 where someone has combined multiple graphics around a particular event:

enter image description here

I accomplish this with arrangeGrob, grid.arrange or par(mfrow=c()) + layout (sometimes) in R.

I'm inclined to call this a dashboard as an info-graphic could be this type of display but an info-graphic may only have a single graph (I tend to think most all graphs are info-graphs).

This web site says that Stephen Few says this about dashboards:

A dashboard is a visual display of the most important information needed to achieve one or more objectives; consolidated and arranged on a single screen so the information can be monitored at a glance.

This seems closest of the terms I'm throwing around but am uncertain. It implies that the display should be on a computer. My description is not limited to computer, for example I could display this on printed paper in a magazine. Perhaps there's a better term(s) than dashboard.

What is the term(s) used for combining different type (i.e., not faceted), yet related, graphs into a single display?

$\endgroup$
  • 2
    $\begingroup$ How about "ensemble"? In music, that's a bunch of people who play together, so why couldn't it be a bunch of graphs that play together? To me, "dashboard" sounds more like some kind of control panel. $\endgroup$ – Russ Lenth Aug 31 '14 at 2:42
  • 2
    $\begingroup$ I don't think there is a distinctive term here; nor is one needed. I agree that (e.g.) "small multiples" usually refers to repetitions of the same design and that your example is not lattice or trellis graphics. In my view any simple wording describing use of a combination is adequate. Indeed people often just write things like "Figure 7 shows a scatter plot, histograms of the distributions, a quantile-normal plot of residuals and leverage information" without needing to emphasise that different graphs are being combined. $\endgroup$ – Nick Cox Aug 31 '14 at 10:16
  • $\begingroup$ Nick I would agree but I'm writing a dissertation and arguing the the interaction between graphs types layers meaning, so here I need to call it something. At the moment Bertin's demonstrational unit is looking most promising. $\endgroup$ – Tyler Rinker Sep 1 '14 at 0:38
  • $\begingroup$ I don't follow your logic. Writing a dissertation, like anything else, doesn't imply that you need a different term for everything different. See also below. $\endgroup$ – Nick Cox Sep 1 '14 at 8:45
  • $\begingroup$ To paraphrase Freud: Sometimes a combined graph is just a combined graph. $\endgroup$ – Nick Cox Sep 1 '14 at 8:46
6
$\begingroup$

Stephen Few (unsuccessfully) addressed this issue in a 2007 article, Dashboard Confusion Revisited and suggested the term "faceted analytical display":

A “faceted analytical display” is a set of interactive charts (primarily graphs and tables) that simultaneously reside on a single screen, each of which presents a somewhat different view of a common dataset, and is used to analyze that information.

which he differentiates from "dashboards":

The greatest clarification that is needed today is a distinction between dashboards, which are used for monitoring what's going on, and displays that combine several charts on a screen for the purpose of analysis.

However, the distinction hasn't been embraced. I most often hear "dashboard" or something generically descriptive like "set of graphs" for such displays.

In a 2007 white paper for Tableau Software, Three Blind Men and an Elephant, Few continues to promote "faceted analytical display" while noting that Tableau uses the term "dashboard".

Perhaps the best we can hope for at keeping a distinction is a qualification of "dashboard" such as "static dashboard" or "analytic dashboard". Or a portmanteau like "graphboard".

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Terrific answer. I'm going to hold off a bit on the check to give time for other responses. I email Few for clarification, can faceted analytical display be extended to static? I'll report back here his sentiment. $\endgroup$ – Tyler Rinker Aug 31 '14 at 19:09
  • $\begingroup$ Stephen Few responded and said "I wouldn't use it [faceted analytical display] to describe static views of data". The idea was good but to use it to describe static plots would not be in keeping with Few's intention. I think the generic term you suggest "set of graphs" or Bertin's "demonstrational unit" may be more appropriate. $\endgroup$ – Tyler Rinker Sep 1 '14 at 0:27
  • 1
    $\begingroup$ Nice follow-up. I didn't catch that your question was only about static graphs. $\endgroup$ – xan Sep 1 '14 at 13:26
  • $\begingroup$ Not necessarily static graphs but a term that incorporates static and dynamic combinations of graphs*. $\endgroup$ – Tyler Rinker Sep 1 '14 at 14:25
5
$\begingroup$

As I dug through Bertin (1983) on page 32 he describes a collection of a series of diagrams, tables, maps, etc. as a demonstrational unit.

...the graphic representation will lead to a series of diagrams, double-entry tables..., maps etc...at the moment of publication, the entire set of diagrams must be conceived as a demonstrational unit. The layout must be considered a scientific problem, linked to the imperatives of reading and comparison, before it can be treated as an aesthetic problem. p. 32

Bertin, Jacques. (1983). Semiology of graphics: Diagrams, networks, maps. Madison, WI: University of Wisconsin Press.

| cite | improve this answer | |
$\endgroup$
  • 1
    $\begingroup$ Bertin's graphical ideas were often extremely good, but reading his prose is often like swimming through mud. I don't blame the translations, either. I don't find that term at all evocative and I am not surprised it never caught on. (I have looked through Bertin several times and never thought it worth noting, let alone repeating.) But, but, but: If there is no good term, it's your chance to suggest a better one in your dissertation, at least for your purposes. $\endgroup$ – Nick Cox Sep 1 '14 at 8:44
  • $\begingroup$ @Nick this may be the most sensible. I'm continuing to read around this area and if I find something fitting I'll use that and report, if not this may be the most sensible approach... To paraphrase Freud: Sometimes a combined graph is just a combined graph :-) $\endgroup$ – Tyler Rinker Sep 1 '14 at 14:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.