6
$\begingroup$

I recently updated to Mathematica 11.2, and I very much like the automatic PlotLabels feature that I discovered. However, it picks a somewhat silly ordering when the data I'm plotting cuts off at different x-locations. Say I have data that looks something like this:

slopes = Table[i Round[Abs[Sin[i]], 0.1], {i, 1, 5, 0.2}];
mydata = Table[Select[Table[{x, 1 + slope * x^2}, {x, 1, 10, 0.1}], #1[[2]] < 100 &],
   {slope, slopes}];

If I just ListPlot it, Mathematica does something smart about plot label vertical positioning which I haven't managed to recreate manually:

ListPlot[mydata, PlotLabels -> slopes, ImageSize -> Large]

smart PlotLabels

However, the ordering is all off; I want to order them by the slope, not by the y-value at the location where my data happened to cut off. So I tried using Callout instead, with a manually specified anchor position:

ListPlot[Map[Callout[#1[[1]], #1[[2]], Automatic, 5] &, 
    Transpose[{mydata, slopes}]], ImageSize -> Large]

This gives the horrendous-looking

bad callout

Okay, fine, I can give it a position manually. But 10 and {10, After} both give identical, very bad results:

manual x position


What I want, from most important to least important (but preferably all of them):

  1. the labels should not overlap and should not be cut off
  2. the labels should have a pointer to some point on the curve
  3. the labels should be in slope-order, or, more precisely, the order that the data has at a particular x-position that I specify
  4. the labels should all be along the right side of the plot, the axis should not extend past the end of the data

How can I accomplish this?

$\endgroup$
2

1 Answer 1

2
$\begingroup$

Using Callout with explicit positions for labels in the setting of PlotLabels combined with manual reordering of input data:

slopes = Sort@Table[i Round[Abs[Sin[i]], 0.1], {i, 1, 5, 0.2}];
mydata = Table[Select[Table[{x, 1 + slope * x^2}, {x, 1, 10, 0.1}], #1[[2]] <= 100 &],
   {slope, slopes}];

ordering ={1,2,21,20,19,18,17,16,15,14,13,11,12,3,4,10,9,8,7,6,5};
mydataordered = mydata[[ordering]];
slopesordered = slopes[[ordering]];
xpos = 12;
yrange = {5, 100};

ListPlot[mydataordered,
 PlotLabels -> MapIndexed[Callout[#,{xpos, Rescale[#2[[1]], {1, Length@mydata}, yrange]}, 
     mydataordered[[All,-1]][[#2[[1]]]], 
     LeaderSize -> {{Automatic, Automatic, 0}, {30, 0 Degree}}]&, slopesordered], 
 ImagePadding -> {{Automatic,200}, { Automatic, Automatic}}, 
 ImageSize -> 800, PlotRangeClipping -> False]

enter image description here

Use xpos = 11; yrange={20, 100}; to get

enter image description here

Alternatively, use

ListPlot[MapIndexed[Callout[#, slopesordered[[#2[[1]]]], 
    {xpos, Rescale[#2[[1]], {1, Length @ mydata}, yrange]}, 
     mydataordered[[All,-1]][[#2[[1]]]], 
  LeaderSize -> {{Automatic, Automatic, 0}, {30, 0 Degree}}]&, mydataordered],
ImagePadding -> {{Automatic, 200}, {Automatic, Automatic}}, 
ImageSize -> 800, PlotRangeClipping -> False]

to get the same result as above.

Update: An alternative layout for the Callouts:

combined = Transpose[{mydata, slopes}];
{top, left} = Pick[combined, #[[1, -1, 1]]  < 9.9 & /@ combined, #]&/@{True, False};
{datatop, slopestop} = Transpose[SortBy[top, #[[1,-1,1]]&]];
{dataleft, slopesleft} = Transpose[SortBy[left, #[[1,-1,-1]]&]];
dL = MapIndexed[Callout[#, slopesleft[[#2[[1]]]], 
 {11, Rescale[#2[[1]], {1, Length @ dataleft}, {30, 90}]}, dataleft[[All,-1]][[#2[[1]]]], 
 LeaderSize -> {{Automatic, Automatic, 0}, {10, 0 Degree}}]&, dataleft];
dT = MapIndexed[Callout[#, slopestop[[#2[[1]]]], 
 {Rescale[#2[[1]],{1, Length @ datatop}, {1, 12}], 110}, datatop[[All, -1]][[#2[[1]]]], 
 LeaderSize -> {{Automatic, Automatic, 0}, {10, 90 Degree}}]&, datatop];

ListPlot[Join[dL,dT], ImagePadding -> {{Automatic,200},{ Automatic, 100}}, 
 ImageSize->800, PlotRangeClipping->False]

enter image description here

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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