Insert vertical space in a heatmap

[Mario Liebensteiner]

Dear Stata users,

I am trying to create a heatmap and would like to include vertical spacing on the y-axis. However, I have been unable to achieve this. I would greatly appreciate any advice or suggestions on how to accomplish this.

To illustrate what I am aiming for, I have included a picture that I produced in PowerPoint as an example of how it should look. The data and code is provided below.

Thank you in advance for your help!

Best,
Mario Liebensteiner



I am using this example dataset:

country value sort_h
FR .66 3
FR .55 2
FR .22 1
AT .77 3
AT .44 2
AT .11 1

My code is:

heatplot value i.sort_h country, ///
cuts(0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1) keylabels(, interval) ///
yscale(noline) ///
ylabel( 1 "ref" 2 "value 1" 3 "value 2" , nogrid) ///
xlabel(, nogrid) ///
color(inferno, reverse) ///
xtitle("") ytitle("")

[Nick Cox]

This is very likely to be annoying as well as irrelevant to the question, but a heat map for six values of a numeric outcome for two categorical predictors?

I recommend strongly a simple graph or dot chart.

[Mario Liebensteiner]

I have 24*28 = 672 datapoint per heatmap. This here is just for simplicity.

[Nick Cox]

Understood, but we never know your real question unless and until you tell us what it is.

I think a heatmap for such data is unlikely to be helpful, but then the objections are different, At a minimum row and column order need optimisation. Each tile will be very small and you want to squeeze them further.

I have had not too bad results for multiple bar charts of similar complexity, but it's usually a struggle, whatever you do.

[Mario Liebensteiner]

Nick, thank you for pointing this out. I just wanted to keep it as simple as possible.


To everybody: I managed to achieve what I wanted. The trick was simple, but it took me long to find out. I simply added a missing-value entry for each country where the sorting variable had a value of 2.


To provide some context, I use the heat maps to show and compare estimates of marginal and average emissions factors for different electricity markets in Europe. Here are the heat maps:

[Nick Cox]

Now that time of day has been revealed as one predictor, I would want to try time series plots and in particular front-and-back plots. fabplot is a suitable search term for the forum.

[Mario Liebensteiner]

Thank you indeed for the hint! I'll see into it. If I am able to produce a nice graph, I will post it here.