Creating a bar chart with error when all variables are categorical

[Jack Prado]

Hi everyone,

I am trying to create bar graphs with error bars. My data consists of an individual id called id, place they are moving called destination, and an individual's race called race. I also have dummy variables for each of the possible destinations. My data is below.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(id destination race) byte(east west south north)
 1 2 4 0 1 0 0
 2 3 1 0 0 1 0
 3 4 1 0 0 0 1
 4 1 4 1 0 0 0
 5 2 5 0 1 0 0
 6 3 1 0 0 1 0
 7 3 4 0 0 1 0
 8 3 1 0 0 1 0
 9 2 5 0 1 0 0
10 3 5 0 0 1 0
11 4 2 0 0 0 1
12 2 1 0 1 0 0
13 1 4 1 0 0 0
14 1 1 1 0 0 0
15 1 1 1 0 0 0
16 1 5 1 0 0 0
17 2 2 0 1 0 0
18 3 4 0 0 1 0
19 1 5 1 0 0 0
20 1 5 1 0 0 0
21 4 3 0 0 0 1
22 1 5 1 0 0 0
23 3 1 0 0 1 0
24 2 4 0 1 0 0
25 3 2 0 0 1 0
26 1 2 1 0 0 0
27 3 1 0 0 1 0
28 1 4 1 0 0 0
29 3 3 0 0 1 0
30 3 4 0 0 1 0
31 2 4 0 1 0 0
32 2 4 0 1 0 0
33 2 2 0 1 0 0
34 2 4 0 1 0 0
35 1 4 1 0 0 0
36 2 1 0 1 0 0
37 3 2 0 0 1 0
38 1 2 1 0 0 0
39 3 1 0 0 1 0
40 3 5 0 0 1 0
end

label values destination area
label def area 1 "east", modify
label def area 2 "west", modify
label def area 3 "south", modify
label def area 4 "north", modify
label values race Race
label def Race 1 "African American", modify
label def Race 2 "Asian", modify
label def Race 3 "Latino", modify
label def Race 4 "White", modify
label def Race 5 "Other", modify

I subsetted each racial group and used the graph command to chart where individuals are planning to move. After subsetting for the five racial groups, I used the user package grc1leg2 to combine them but there isn't an option to add confidence intervals.

*Subset for African American
preserve
keep if race==1

**Original destination for African American
graph hbar east west south north, over(race, label(nolabel)) ///
legend(size(small) cols(5) region(lstyle(none)) ///
label(1 "east") label(2 "west") label(3 "south") label(4 "north")) ///
title("African American") ///
ylabel(0 .1 .2 .3 .4 .5 .6 .7, angle(horizontal)) ///
ytitle("Percentage") ///
blabel(bar, format(%4.2f)) ///
saving("AA", replace)
restore

*Subset for Asian
preserve
keep if race==2

*Transfer to segment disaggregated by race
graph hbar east west south north, over(race, label(nolabel)) ///
legend(size(small) cols(5) region(lstyle(none)) ///
label(1 "east") label(2 "west") label(3 "south") label(4 "north")) ///
title("Asian") ///
ylabel(0 .1 .2 .3 .4 .5 .6 .7, angle(horizontal)) ///
ytitle("Percentage") ///
blabel(bar, format(%4.2f)) ///
saving("Asian", replace)
restore


*Subset for Latino
preserve
keep if race==3

*Original destination for Latio
graph hbar east west south north, over(race, label(nolabel)) ///
legend(size(small) cols(5) region(lstyle(none)) ///
label(1 "east") label(2 "west") label(3 "south") label(4 "north")) ///
title("Latino") ///
ylabel(0 .1 .2 .3 .4 .5 .6 .7, angle(horizontal)) ///
ytitle("Percentage") ///
blabel(bar, format(%4.2f)) ///
saving("Latino", replace)
restore


*Subset for White
preserve
keep if race==4

*Original destination for White
graph hbar east west south north, over(race, label(nolabel)) ///
legend(size(small) cols(5) region(lstyle(none)) ///
label(1 "east") label(2 "west") label(3 "south") label(4 "north")) ///
title("White") ///
ylabel(0 .1 .2 .3 .4 .5 .6 .7, angle(horizontal)) ///
ytitle("Percentage") ///
blabel(bar, format(%4.2f)) ///
saving("White", replace)
restore


*Subset for Other
preserve
keep if race==1

*Original destination for Other
graph hbar east west south north, over(race, label(nolabel)) ///
legend(size(small) cols(5) region(lstyle(none)) ///
label(1 "east") label(2 "west") label(3 "south") label(4 "north")) ///
title("Other") ///
ylabel(0 .1 .2 .3 .4 .5 .6 .7, angle(horizontal)) ///
ytitle("Percentage") ///
blabel(bar, format(%4.2f)) ///
saving("Other", replace)
restore

*Combine all the charts together
grc1leg2 AA.gph Asian.gph Latino.gph White.gph Other.gph,  ycommon

As a result, I am using the following website to help me estimate the error bars.

https://stats.oarc.ucla.edu/stata/fa...th-error-bars/

I tried the following code but this is not the same as the website.

collapse (mean) meandestination= destination (sd) sddestination=destination (count) n=destination, by(race)

I am not sure how to collapse the data the way the website suggests. Any suggestions would be greatly appreciated. Thanks.