Stacking a bar graph

[Rosa Blau]

Hi,

I have a binary variable (childcare_nofsa) and a categorical variable (state). I'd like to produce a stacked bar graph showing the frequencies of "provided" and "not provided" response categories in childcare_nofsa per state.

Here are data examples for the two variables:

childcare_nofsa

* Example generated by -dataex-. To install: ssc install dataex
clear
input float childcare_nofsa
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
.
0
0
0
0
0
0
0
0
.
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
end
label values childcare_nofsa childcare_nofsa
label def childcare_nofsa 0 "Not provided", modify
label def childcare_nofsa 1 "Provided", modify

* Example generated by -dataex-. To install: ssc install dataex
clear
input float state
1
1
1
1
1
2
2
2
2
2
2
2
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
end
label values state state
label def state 1 "AK", modify
label def state 2 "AL", modify
label def state 3 "AR", modify
label def state 4 "AZ", modify
label def state 5 "CA", modify

So far, I've used the following code

graph bar (count) childcare_nofsa, over(state) stack

That gets me the digital picture I've inserted here.

What I'd like to achieve however is
(1) Stacking the responses in terms of the relative frequencies of the values of 1 and 0 in "childcare_nofsa"
(2) Getting the states names to appear vertically so they are all individually legible.
--> I've added an analog picture of what's in my mind's eye.

Grateful for any help

[Nick Cox]

Please give a data example including both your variables.

You may need to re-think your design any way. Minimally, you have 50 states (and what about DC, Puerto Rico, Guam, and so forth?). Will your design really be readable even with vertical labels? What about a horizontal chart?

Isn't it predictable that CA NY TX and so on will just dwarf WY and so on?

[Rosa Blau]

Thanks for your response Nick.

I gave data examples of each variable, but did you mean together?

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(childcare_nofsa state)
0 1
0 1
0 1
0 1
1 1
0 2
0 2
0 2
0 2
0 2
0 2
0 2
0 3
0 3
1 3
0 3
0 3
0 3
0 4
0 4
1 4
0 4
0 4
0 4
1 4
0 4
0 4
0 4
0 4
0 4
0 5
1 5
0 5
0 5
1 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
1 5
0 5
0 5
0 5
1 5
1 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
. 5
0 5
0 5
0 5
0 5
0 5
. 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
1 5
0 5
0 5
0 5
0 5
1 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
0 5
end
label values childcare_nofsa childcare_nofsa
label def childcare_nofsa 0 "Not provided", modify
label def childcare_nofsa 1 "Provided", modify
label values state state
label def state 1 "AK", modify
label def state 2 "AL", modify
label def state 3 "AR", modify
label def state 4 "AZ", modify
label def state 5 "CA", modify

Regarding chart design: Indeed, I have the 50 states + DC. And exactly those three states do dwarf the others. But I think I need the graph to make a larger point about my analysis. I also may end up only constructing a graph with a smaller selection of states, but for that reason too it would be great to know how.

[Nick Cox]

Yes, thanks; that is the kind of data example I was seeking.

graph bar and graph hbar can accommodate what you're asking for, at least since Stata 13 as updated on 9 October 2014, but I remain more familiar with the convenience wrapper catplot from SSC I first wrote in 2003 and revised in 2010. .

Here is a stab at a horizontal bar chart for your example.

bysort state : gen total = _N 

set scheme s1color 

* ssc install catplot 
catplot childcare_nofsa state, asyvars stack var2opts(sort(total) descending) bar(1, bfcolor(red*0.2) blcolor(red)) bar(2, bfcolor(blue*0.2) blcolor(blue))

[Rosa Blau]

Thanks a lot Nick! This is what I've got now. Do you know how I could space the bars so the labels of the states don't overlap?

[Nick Cox]

This is precisely the problem warned about in #2. So you can make your text smaller, change the aspect ratio, make your graph larger or recast in two panels as explained in section 6 of https://www.stata-journal.com/articl...article=gr0080

To get more specific help, please show the results of

contract state childcare_nofsa
dataex

[Rosa Blau]

Ah, when you asked "will the design really be readable" I thought you meant having to read states names vertically. I didn't realise they would be illegible because they might physically overlap. Anyway, here is the data you requested:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(childcare_nofsa state) byte _freq
0  1  4
1  1  1
0  2  7
0  3  5
1  3  1
0  4 10
1  4  2
0  5 67
1  5  8
.  5  2
0  6 19
1  6  3
0  7  9
1  7  1
0  8  2
1  8  1
0  9  5
0 10 26
1 10  8
0 11 15
1 11  6
0 12  2
0 13 20
1 13  1
0 14  5
0 15 42
1 15  9
0 16 17
1 16  2
0 17 15
1 17  2
0 18  8
0 19  8
1 19  2
0 20 23
1 20  2
0 21 14
1 21  2
0 22 10
0 23 28
1 23  3
0 24 33
1 24  2
0 25  9
1 25  9
0 26  4
1 26  1
0 27  3
1 27  2
0 28 11
1 28  2
0 29  3
0 30  4
1 30  2
0 31  7
1 31  1
0 32 22
1 32  5
. 32  1
0 33  3
1 33  1
0 34  5
0 35 32
1 35  7
0 36 30
1 36  6
0 37  7
0 38 13
1 38  3
0 39 35
1 39  5
. 39  2
0 40  1
1 40  1
0 41  8
1 41  3
0 42  6
1 42  1
0 43 13
1 43  2
. 43  1
0 44 57
1 44  5
0 45  7
1 45  1
0 46 30
1 46  5
. 46  1
0 47  2
1 47  1
0 48 18
1 48  3
0 49 30
1 49  4
0 50  3
1 50  1
1 51  2
0  . 16
1  .  8
.  .  2
end
label values childcare_nofsa childcare_nofsa
label def childcare_nofsa 0 "Not provided", modify
label def childcare_nofsa 1 "Provided", modify
label values state state
label def state 1 "AK", modify
label def state 2 "AL", modify
label def state 3 "AR", modify
label def state 4 "AZ", modify
label def state 5 "CA", modify
label def state 6 "CO", modify
label def state 7 "CT", modify
label def state 8 "DC", modify
label def state 9 "DE", modify
label def state 10 "FL", modify
label def state 11 "GA", modify
label def state 12 "HI", modify
label def state 13 "IA", modify
label def state 14 "ID", modify
label def state 15 "IL", modify
label def state 16 "IN", modify
label def state 17 "KS", modify
label def state 18 "KY", modify
label def state 19 "LA", modify
label def state 20 "MA", modify
label def state 21 "MD", modify
label def state 22 "ME", modify
label def state 23 "MI", modify
label def state 24 "MN", modify
label def state 25 "MO", modify
label def state 26 "MS", modify
label def state 27 "MT", modify
label def state 28 "NC", modify
label def state 29 "ND", modify
label def state 30 "NE", modify
label def state 31 "NH", modify
label def state 32 "NJ", modify
label def state 33 "NM", modify
label def state 34 "NV", modify
label def state 35 "NY", modify
label def state 36 "OH", modify
label def state 37 "OK", modify
label def state 38 "OR", modify
label def state 39 "PA", modify
label def state 40 "RI", modify
label def state 41 "SC", modify
label def state 42 "SD", modify
label def state 43 "TN", modify
label def state 44 "TX", modify
label def state 45 "UT", modify
label def state 46 "VA", modify
label def state 47 "VT", modify
label def state 48 "WA", modify
label def state 49 "WI", modify
label def state 50 "WV", modify
label def state 51 "WY", modify

Thanks again for taking the time to walk be through this.

[Nick Cox]

Here's some technique with your data example repeated for convenience.

In terms of

you can make your text smaller, change the aspect ratio, make your graph larger or recast in two panels as explained in section 6 of https://www.stata-journal.com/articl...article=gr0080

this uses only the last trick, so

make your text smaller, change the aspect ratio, make your graph larger

remain as possibilities.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(childcare_nofsa state) byte _freq
0  1  4
1  1  1
0  2  7
0  3  5
1  3  1
0  4 10
1  4  2
0  5 67
1  5  8
.  5  2
0  6 19
1  6  3
0  7  9
1  7  1
0  8  2
1  8  1
0  9  5
0 10 26
1 10  8
0 11 15
1 11  6
0 12  2
0 13 20
1 13  1
0 14  5
0 15 42
1 15  9
0 16 17
1 16  2
0 17 15
1 17  2
0 18  8
0 19  8
1 19  2
0 20 23
1 20  2
0 21 14
1 21  2
0 22 10
0 23 28
1 23  3
0 24 33
1 24  2
0 25  9
1 25  9
0 26  4
1 26  1
0 27  3
1 27  2
0 28 11
1 28  2
0 29  3
0 30  4
1 30  2
0 31  7
1 31  1
0 32 22
1 32  5
. 32  1
0 33  3
1 33  1
0 34  5
0 35 32
1 35  7
0 36 30
1 36  6
0 37  7
0 38 13
1 38  3
0 39 35
1 39  5
. 39  2
0 40  1
1 40  1
0 41  8
1 41  3
0 42  6
1 42  1
0 43 13
1 43  2
. 43  1
0 44 57
1 44  5
0 45  7
1 45  1
0 46 30
1 46  5
. 46  1
0 47  2
1 47  1
0 48 18
1 48  3
0 49 30
1 49  4
0 50  3
1 50  1
1 51  2
0  . 16
1  .  8
.  .  2
end
label values childcare_nofsa childcare_nofsa
label def childcare_nofsa 0 "Not provided", modify
label def childcare_nofsa 1 "Provided", modify
label values state state
label def state 1 "AK", modify
label def state 2 "AL", modify
label def state 3 "AR", modify
label def state 4 "AZ", modify
label def state 5 "CA", modify
label def state 6 "CO", modify
label def state 7 "CT", modify
label def state 8 "DC", modify
label def state 9 "DE", modify
label def state 10 "FL", modify
label def state 11 "GA", modify
label def state 12 "HI", modify
label def state 13 "IA", modify
label def state 14 "ID", modify
label def state 15 "IL", modify
label def state 16 "IN", modify
label def state 17 "KS", modify
label def state 18 "KY", modify
label def state 19 "LA", modify
label def state 20 "MA", modify
label def state 21 "MD", modify
label def state 22 "ME", modify
label def state 23 "MI", modify
label def state 24 "MN", modify
label def state 25 "MO", modify
label def state 26 "MS", modify
label def state 27 "MT", modify
label def state 28 "NC", modify
label def state 29 "ND", modify
label def state 30 "NE", modify
label def state 31 "NH", modify
label def state 32 "NJ", modify
label def state 33 "NM", modify
label def state 34 "NV", modify
label def state 35 "NY", modify
label def state 36 "OH", modify
label def state 37 "OK", modify
label def state 38 "OR", modify
label def state 39 "PA", modify
label def state 40 "RI", modify
label def state 41 "SC", modify
label def state 42 "SD", modify
label def state 43 "TN", modify
label def state 44 "TX", modify
label def state 45 "UT", modify
label def state 46 "VA", modify
label def state 47 "VT", modify
label def state 48 "WA", modify
label def state 49 "WI", modify
label def state 50 "WV", modify
label def state 51 "WY", modify

set scheme s1color  

gen which = state > 26 
rename _freq freq 
egen total = total(-freq), by(state)

local colors bar(1, fcolor(red*0.2) lcolor(red)) bar(2, fcolor(blue*0.2) lcolor(blue))


graph hbar (count) [fw=freq] , over(childcare_nofsa) over(state) by(which, note("")) nofill stack asyvars subtitle("", fcolor(none)) name(G1, replace) ///
`colors'

egen group = group(total state) 
labmask group, values(state) decode 
replace which = group > 26 

graph hbar (count) [fw=freq] , over(childcare_nofsa) over(group) by(which, note("")) nofill stack asyvars subtitle("", fcolor(none)) name(G2, replace) ///
`colors'

[Rosa Blau]

Thanks a lot Nick! This is very helpful and I think I can tweak it from here.