How do I create multiple group scatter plots with the same legend?

I have a table with outputs from multiple model Runs. I ran this model for 24 rivers, with different input datasets. Now I want to create a scatterplot, showing the range of values for each river, by plotting the minimum and maximum value for each river. I achieved this, but now I want to show in which run each of these minimum or maximum values was produced. I wanted to do this using gscatter, and using the Run Name as group variable. However, because I am using gscatter twice (once for the minimum values, and then again for the maximum values), and the Run Names differ for each scatterplot (the minimum values were often produced by different Runs than the maximum values), the legend only shows the colors used for the last scatterplot (for the maximums) that was created, and thus doesn't show the correct values for the first scatterplot (the minimums). Below is the code I am using

X=outputtable.Orig;

Y1=min([outputtable{:,2:9},outputtable{:,11:12}]')'; %find the minimum values for the relevant columns

Y2=max([outputtable{:,2:9},outputtable{:,11:12}]')'; %find the maximum values for the relevant columns

for i=1:height(outputtable)

minloc(i)=find(outputtable{i,2:13}==Y1(i))+1; %find the column in which the minimum value is located, using +1 because of table indices {:,2:13}

maxloc(i)=find(outputtable{i,2:13}==Y2(i))+1; %find the column in which the maximum value is located, using +1 because of table indices {:,2:13}

end

minvar=string(outputtable.Properties.VariableNames(minloc)); %find the variable name (=Run Name) corresponding to the minimum value per river

maxvar=string(outputtable.Properties.VariableNames(maxloc)); %find the variable name (=Run Name) corresponding to the maximum value per river

gscatter(X,Y1,outputtable.MinVar)

hold on

gscatter(X,Y2,outputtable.MaxVar)

Answer 1

[Varun]

Hello!

As per my understanding, you’re facing an error with the auto-generated legend for the plots generated on the same axes. I tried plotting similar plots using some sample data from a MATLAB example. This is what helped me fix the issue:

1. You can pop-out the figure into a figure window and manually toggle (turn off and on) the legend once. The refreshed legend will show up with information regarding both plots in the same legend box. However, this does not reflect on the plot in the live editor.

2. You can specify the legend in the code itself. Since you are using the variables “outputtable.MinVar” and “outputtable.MaxVar” as the grouping variables for the two plots, their unique values will show up in the legend. So, this code snippet should do the job:

    legend(string([unique(outputtable.MinVar), unique(outputtable.MaxVar)]))

This will add the required legend to the plot. Hope this helps!