Different colors for two groups of bars in a bar chart

[Zan Strabac]

Dear all,
I am trying to make a bar chart that shows mean value of a variable (attitudes toward immigrants) for 19 European countries. I want to use different colors (or intensities of same color) for East European (ex-communist) countries and West European countries. I have created a dummy variable that distinguishes between these two categories of countries, but was not able to find any info about how i can make the graph.
Syntax is bellow and an example of the graph using only one color for the bars is attached.

Any help is much appreciated.
Best,
Zan Strabac

graph bar (mean) immigr_att [pw=pspwght], over(country, sort(1) lab(angle(45))) intensity (50)

[Nick Cox]

Please give a data example and show your existing graph as .png not .gph. FAQ Advice #12 explains both points.

[Zan Strabac]

Please give a data example and show your existing graph as .png not .gph. FAQ Advice #12 explains both points.

Ok, trying to post in accordance with FAQ #12:

 graph bar (mean) immigr_att [pw=pspwght], over(country, sort(1) lab(angle(45))) intensity (50)

Presenting a short random sample of data, and only 5 countries using randomtag and dataex:

* Example generated by -dataex-. For more info, type help dataex
clear
input float(immigr_att pspwght) long(country excom)
        5 1.2814763 1 1
        4  .4017383 1 1
4.6666665  .6357259 1 1
       10  .6878853 1 1
4.3333335  .6646395 1 1
4.6666665  .9279124 1 1
        7  .7465485 1 1
 6.333333  .9133835 1 1
        .  2.667386 1 1
        .  .6074823 1 1
1.6666666 1.8672407 1 1
 3.333333  .4003073 1 1
 7.333333   .872346 2 0
        8 1.1248171 2 0
        9 1.0638217 2 0
2.3333333  .8510109 2 0
 8.666667  .9286219 2 0
        7 1.1564138 2 0
 8.333333 1.0456595 2 0
        .  1.227084 2 0
 3.666667  .9934008 3 1
        2  .5836357 3 1
        6  .9704045 3 1
 5.333333  .6254382 3 1
        .  .6605008 3 1
        0  .6483748 3 1
1.6666666 1.1796651 3 1
 6.666667   .907792 3 1
        6   .842384 4 1
 3.333333  .7788791 4 1
        7 1.2844403 4 1
 3.333333 1.1847209 4 1
 3.666667  .7790892 4 1
1.6666666  .9848319 4 1
        5  .7565995 4 1
2.3333333 1.2162174 4 1
        .  1.331813 4 1
        6  .8426083 4 1
        5  .7302645 4 1
 6.666667 1.2549235 5 0
        7  .7530739 5 0
 7.666667  .7660794 5 0
 5.333333  .6529269 5 0
        8  .6093153 5 0
 7.666667  .7660794 5 0
        7  .5989709 5 0
 5.666667 1.0349259 5 0
 6.333333 1.2931395 5 0
        6  .9268494 5 0
 2.666667 1.3568442 5 0
end
label values country cntr_lbl
label def cntr_lbl 1 "Bulgaria", modify
label def cntr_lbl 2 "Switzerland", modify
label def cntr_lbl 3 "Czechia", modify
label def cntr_lbl 4 "Estonia", modify
label def cntr_lbl 5 "Finland", modify

And finally, attaching the original graph (that i would like to createin separate bar colors for East and West Eu. countries) as a .png file:

[Nick Cox]

Thanks for the extra detail. Try

separate immigr_att, by(excom)
graph bar (mean) immigr_att? [pw=pspwght], over(country, sort(1) lab(angle(45))) intensity(50) nofill legend(order(1 "West European" 2 "East European")) ytitle(Some wording here)

except that I recommend hbar not least to avoid those 45 degree slopes (steeper than scree or talus!).

For more detail, see https://www.stata-journal.com/articl...article=gr0049

https://en.wikipedia.org/wiki/Scree

[Zan Strabac]

Thanks for the code Nick, but I was trying to use different colors for East and West, while preserving the ordering by the dependent variable (sorry for not mentioning explicitly in the first post). So, basically a version of what have been done for Missisippi in the River dataset in your article in Stata Journal. I have tried to adapt the code from the article, but was not able to make it work.

[Nick Cox]

* Example generated by -dataex-. For more info, type help dataex
clear
input float(immigr_att pspwght) long(country excom)
        5 1.2814763 1 1
        4  .4017383 1 1
4.6666665  .6357259 1 1
       10  .6878853 1 1
4.3333335  .6646395 1 1
4.6666665  .9279124 1 1
        7  .7465485 1 1
 6.333333  .9133835 1 1
        .  2.667386 1 1
        .  .6074823 1 1
1.6666666 1.8672407 1 1
 3.333333  .4003073 1 1
 7.333333   .872346 2 0
        8 1.1248171 2 0
        9 1.0638217 2 0
2.3333333  .8510109 2 0
 8.666667  .9286219 2 0
        7 1.1564138 2 0
 8.333333 1.0456595 2 0
        .  1.227084 2 0
 3.666667  .9934008 3 1
        2  .5836357 3 1
        6  .9704045 3 1
 5.333333  .6254382 3 1
        .  .6605008 3 1
        0  .6483748 3 1
1.6666666 1.1796651 3 1
 6.666667   .907792 3 1
        6   .842384 4 1
 3.333333  .7788791 4 1
        7 1.2844403 4 1
 3.333333 1.1847209 4 1
 3.666667  .7790892 4 1
1.6666666  .9848319 4 1
        5  .7565995 4 1
2.3333333 1.2162174 4 1
        .  1.331813 4 1
        6  .8426083 4 1
        5  .7302645 4 1
 6.666667 1.2549235 5 0
        7  .7530739 5 0
 7.666667  .7660794 5 0
 5.333333  .6529269 5 0
        8  .6093153 5 0
 7.666667  .7660794 5 0
        7  .5989709 5 0
 5.666667 1.0349259 5 0
 6.333333 1.2931395 5 0
        6  .9268494 5 0
 2.666667 1.3568442 5 0
end
label values country cntr_lbl
label def cntr_lbl 1 "Bulgaria", modify
label def cntr_lbl 2 "Switzerland", modify
label def cntr_lbl 3 "Czechia", modify
label def cntr_lbl 4 "Estonia", modify
label def cntr_lbl 5 "Finland", modify

separate immigr_att, by(excom)
graph bar (mean) immigr_att? [pw=pspwght], over(country, sort(immigr_att) lab(angle(45))) intensity(50) nofill legend(order(1 "West European" 2 "East European")) ytitle(Some wording here) scheme(s1color)

[Zan Strabac]

This did the deed, thank you so much, Nick. (And thanx for proposing different colors, using Stata's default blue and red might have some unwanted connotations in this case :-))
I am posting the final graph for the benefit of possible future readers of this tread: