# graphing: show a distant cluster while minimizing negative space?

I am preparing a map of cluster assignments for a publication that seeks to minimize negative space in their figures. This is what my map looks like:

Notice the distant island in orange? I want to emphasize the fact that its far away without showing the all the negative space. The issue in bar/scatter/line graphs is simple: make the axes discontinuous with a zigzag or a gap in the axis line. But for a map like this, where there are no axis and distance is purely qualitative, how do I demonstrate the difference of the orange cluster without the negative space?

• Can you give more context as to what this is used for? It'd help making a decision as to what would work much easier – Zach Saucier Jun 20 '16 at 6:54
• These are cluster assignments, meaning distance is a function of similarity: points in a cluster (or island) are more alike than points belonging to a distant cluster. I want to show that the orange cluster use really different without showing much white space. – Thomas Matthew Jun 20 '16 at 16:53

I know this is a very old question, but its an interesting problem.

here's a solution I came up with. Keep in mind I don't have a full grasp of usage. My goal was the indicate distance in some way visually while reducing the amount of white space overall in the image...

The reduced background image allows for indication of a more accurate distance. While the larger callouts would allow for detailed objects themselves.

I have no clue if this would work in your particular situation. You mention "map" and "island" in the question so those were somewhat at the back of my mind.

If the axiis really don't matter, the obvious choice is to reduce the size of the bounding box (border added for clarity):

• It is probably a matrix having the thing plotted in the diagonal has a specific meaning to the audience if this is the case – joojaa May 21 '16 at 9:30

The clusters appear to be roughly aligned along a 45° axis.

To "compress" the distance between the clusters and preserve their size and relative orientation, position your "zig-zag" break perpendicular to the axis formed by your cluster plot points.

Remove as much of the negative space along the axis as you wish to illustrate your point.