Is it possible to alternate bar graph labels so that they don't overlap?

[Zach Goldberg]

Greetings,

I'm running Stata 15.1 on OSX. I'm trying to graph longitudinal data over three different over() variables (ideology, age group and year). This naturally entails a lot of bars and, consequently, a lot of overlapping value labels. I could reduce the size of the bar labels and/or increase the size of the graph (e.g. xsize(15)), but this would make them harder to read when pasted into a word document. I know you can alternate the y and x labels...is it possible to do the same with bar labels? If not, what's my best option here?.

Here is my syntax:
[CODE][graph bar W2MI if white==1 & W2MI!=. [pweight=weight], over(ideo3) over(age_cat) over(YEAR) asyvar blabel(total, format(%9.2f) size(medsmall)) bargap(30) xsize(12) ytitle("Percent") plotregion(fcolor(white)) graphregion(fcolor(white)) legend(col(1)) title("By Ideology")/CODE]

And this is the graph it produces:

Thanks in advance for your time!

[Nick Cox]

Consider graph hbar or even tabplot (Stata Journal; https://www.statalist.org/forums/for...updated-on-ssc)

If you post the data, people will be able to post alternative suggestions.

collapse W2MI if white==1 [pweight=weight], by(ideo3 age_cat YEAR)
dataex

[Zach Goldberg]

That's the one thing I forgot to include. Here it is:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(W2MI YEAR white) double ideo3 float age_cat
. 1988 1 3 5
. 1988 1 3 1
. 1988 1 3 1
. 1988 1 2 4
. 1988 0 2 6
. 1988 0 3 6
. 1988 1 2 1
. 1988 1 3 1
. 1988 1 3 6
. 1988 1 3 6
. 1988 1 3 2
. 1988 1 1 4
. 1988 1 2 2
. 1988 1 3 5
. 1988 1 3 4
. 1988 1 3 2
. 1988 1 3 6
. 1988 1 3 3
. 1988 1 2 2
. 1988 1 2 5
. 1988 0 1 2
. 1988 1 3 3
. 1988 0 3 6
. 1988 0 3 6
. 1988 0 3 1
. 1988 0 3 3
. 1988 0 3 2
. 1988 1 3 4
. 1988 1 3 2
. 1988 1 3 5
. 1988 0 2 4
. 1988 0 3 5
. 1988 1 1 6
. 1988 1 3 1
. 1988 1 1 5
. 1988 1 2 4
. 1988 1 3 2
. 1988 1 3 6
. 1988 1 3 2
. 1988 1 3 6
. 1988 1 3 3
. 1988 1 2 6
. 1988 1 3 3
. 1988 1 3 3
. 1988 1 3 6
. 1988 1 2 2
. 1988 0 3 4
. 1988 1 1 6
. 1988 1 2 6
. 1988 1 3 6
. 1988 1 2 6
. 1988 1 2 6
. 1988 1 2 5
. 1988 1 3 6
. 1988 1 3 2
. 1988 1 3 3
. 1988 1 1 1
. 1988 0 1 3
. 1988 1 3 6
. 1988 1 3 3
. 1988 1 3 4
. 1988 1 2 4
. 1988 1 1 2
. 1988 1 3 2
. 1988 1 3 6
. 1988 0 2 6
. 1988 1 2 3
. 1988 1 3 3
. 1988 1 3 6
. 1988 1 2 3
. 1988 1 3 6
. 1988 1 2 4
. 1988 1 3 6
. 1988 1 3 5
. 1988 1 2 6
. 1988 1 3 4
. 1988 1 3 2
. 1988 1 3 3
. 1988 1 1 5
. 1988 1 3 5
. 1988 1 3 2
. 1988 0 2 2
. 1988 1 1 1
. 1988 1 3 5
. 1988 1 2 4
. 1988 0 2 3
. 1988 1 3 5
. 1988 1 3 1
. 1988 1 3 4
. 1988 1 2 3
. 1988 1 3 4
. 1988 1 3 3
. 1988 0 3 5
. 1988 1 2 3
. 1988 1 3 1
. 1988 1 3 2
. 1988 1 3 6
. 1988 1 3 2
. 1988 0 3 2
. 1988 1 3 6
end
label values ideo3 ideo3
label def ideo3 1 "Liberal", modify
label def ideo3 2 "Moderate", modify
label def ideo3 3 "Conservative", modify
label values age_cat age_cat
label def age_cat 1 "18-24", modify
label def age_cat 2 "25-34", modify
label def age_cat 3 "35-44", modify
label def age_cat 4 "45-54", modify
label def age_cat 5 "55-64", modify
label def age_cat 6 "65+", modify

[Nick Cox]

Thanks for willingness, but that's not much use for illustration as all the responses are missing. I suggested precise code in #2 to get the 3 x 6 x 3 = 54 means shown in your example.

[Marcos Almeida]

I'm just wondering why not presenting the graph as a single column, I mean, each year would represent a row.