graph hbar - bargap(#) and bar(#, fcolor(.)) do not work

[Marry Lee]

Dear all,

My code is the following:

graph hbar (first) estimates, over(authors_num, descending label(labsize(vsmall)))  blabel(bar) stack  graphregion(margin(-40 2 2 2))   ylabel(0(20)210) scheme(s1color)

My initial graph is horizontal green bars.

I want to change the color of only the 3rd bar and reduce the interval between bars, so I use:

graph hbar (first) estimates, over(authors_num, descending label(labsize(vsmall)))  blabel(bar) stack  graphregion(margin(-40 2 2 2))   ylabel(0(20)210) scheme(s1color)  bar(3, fcolor(red)) bargap(0)

Using bar(3, fcolor(red)) changes the original color of all bars to red, not only the 3rd bar.

Using bargap(0) does not change anything in my graph.

Using

graph hbar (first) estimates, over(authors_num, descending label(labsize(vsmall)))  blabel(bar) stack  graphregion(margin(-40 2 2 2))  ylabel(0(20)210) scheme(s1color)  ascategory asyvars bar(3, fcolor(red)) legend(off)

give me a weird graph that is totaly different from what I want:

Am I doing something wrong? can you help please?

[Nick Cox]

To make progress, I think we need to see a data example:

dataex estimates authors_num  

and indeed to know why you asked to stack bars any way.

[Marry Lee]

Nick Cox Thank you for your answer

* Example generated by -dataex-. To install: ssc install dataex
clear
input double estimates float authors_num
   .3  1
  1.6  2
  1.8  3
  2.1  4
  8.7  5
   11  6
   15  7
16.65  8
  139  9
  142 10
  160 11
  163 12
   20  7
  200  9
end

Yes sorry for the stack, it has no need here.
Please use the following code:

graph hbar (first) estimates, over(authors_num, descending relabel(1 "Author 1" 2 "Author 2" 3 "Author 3" 4 "Author 4" 5 "Author 5" 6 "Author 6" 7 "Author 7" 8 "Author 8" 9 "Author 9" 10 "Author 10" 11 "Author 11" 12 "Author 12") label(labsize(vsmall)))  blabel(bar)    ylabel(0(20)210) scheme(s1color)

[Nick Cox]

Thanks for the example, but why are 7 and 9 repeated?

[Marry Lee]

Nick Cox, this is a dataset about papers and the findings in these papers. So in papers 7 and 9 there are 2 estimates in each one of them that I want to show, but I did not figure out how to do that yet, so I would appreciate it if you have any idea about that.

[Nick Cox]

Fine.

scatter authors_num estimates, ysc(reverse) scheme(s1color) yla(1/12, grid ang(h)) ms(Dh)

Consider also xsc(alt) for reasons given in https://journals.sagepub.com/doi/pdf...867X1201200314

[Marry Lee]

Nick Cox thank you so much, this is really helpful.

[Marry Lee]

Nick Cox, sorry this is me again. The solution with scatter is very good, but I just wanted to know whether it would have been possible to present the two values of estimates for the same paper if I have used graph hbar (for papers 7 and 9). No matter how much I think about it, I can't come up with a solution for that. I thought about creating another variable with the two additional values for 7 and 9, but it did not work. If you have any idea about that, can you please share it with me. Thank you!

[Nick Cox]

My idea is that a bar chart is a very poor solution if you want in effect sometimes to show two bars in the same place, which is precisely why I didn't show you code for a bar chart.

If you really want a bar chart, twoway bar is more flexible than graph hbar, but again I don't reommend that.