Pie Chart

[Maneesh Rana]

graph pie rec if year == 22, over(ec_code)

I am using above command with change of year, but stata producing same graph every time. Why so?

* Example generated by -dataex-. To install: ssc install dataex
clear
input int id byte year int head float(ec_code p_code) long rec
 1 22 85 1 13  282984
 1 23 85 1 13  329073
 1 24 85 1 13  369262
 2 22 86 1 13  349992
 2 23 86 1 13  502859
 2 24 86 1 13  510600
 3 22 86 1 13      -9
 3 23 86 1 13       0
 3 24 86 1 13       0
 4 22 86 1 13       0
 4 23 86 1 13       0
 4 24 86 1 13       0
 5 22 86 1 13     -10
 5 23 86 1 13       0
 5 24 86 1 13       0
 6 22 86 1 13       0
 6 23 86 1 13       0
 6 24 86 1 13       0
 7 22 86 1 13  -18613
 7 23 86 1 13       0
 7 24 86 1 13  -22258
 8 22 86 2 13    7332
 8 23 86 2 13       0
 8 24 86 2 13   11519
 9 22 86 3 13     277
 9 23 86 3 13       0
 9 24 86 3 13     400
10 22 86 3 13   19870
10 23 86 3 13       0
10 24 86 3 13   20469
11 22 86 3 13    -152
11 23 86 3 13       0
11 24 86 3 13       0
12 22 86 1 13  465664
12 23 86 1 13  600000
12 24 86 1 13  720066
13 22 86 1 13 -344000
13 23 86 1 13 -400000
13 24 86 1 13 -484214
14 22 86 1 13   71479
14 23 86 1 13   36775
14 24 86 1 13   80000
15 22 86 2 13      62
15 23 86 2 13       0
15 24 86 2 13       0
16 22 86 1 13   44915
16 23 86 1 13       0
16 24 86 1 13   42184
17 22 86 3 13     529
17 23 86 3 13       0
17 24 86 3 13       0
18 22 20 4 13  298575
18 23 20 4 13  330702
18 24 20 4 13  365509
19 22 21 4 13  293864
19 23 21 4 13  321296
19 24 21 4 13  355590
20 22 23 1 13       9
20 23 23 1 13       9
20 24 23 1 13       8
21 22 28 4 13       2
21 23 28 4 13       0
21 24 28 4 13       2
22 22 29 5 13     736
22 23 29 5 13    2000
22 24 29 5 13    2100
23 22 29 5 13      33
23 23 29 5 13     500
23 24 29 5 13     500
24 22 29 5 13      75
24 23 29 5 13       0
24 24 29 5 13       0
25 22 29 6 13     254
25 23 29 6 13      52
25 24 29 6 13      51
26 22 29 5 13       0
26 23 29 5 13       0
26 24 29 5 13       0
27 22 29 3 13       1
27 23 29 3 13       0
27 24 29 3 13       0
28 22 29 6 13     266
28 23 29 6 13      50
28 24 29 6 13     500
29 22 29 5 13    1491
29 23 29 5 13       0
29 24 29 5 13    1000
30 22 29 5 13      26
30 23 29 5 13      50
30 24 29 5 13      55
31 22 29 3 13     136
31 23 29 3 13     900
31 24 29 3 13     900
32 22 29 3 13       2
32 23 29 3 13       0
32 24 29 3 13       0
33 22 29 6 13       1
33 23 29 6 13       0
33 24 29 6 13       0
34 22 29 5 13      52
end
label values year year_id
label def year_id 22 "AE 2021-22", modify
label def year_id 23 "RE 2022-23", modify
label def year_id 24 "BE 2023-24", modify
label values head rec_id
label def rec_id 20 "Corporation Tax", modify
label def rec_id 21 "Taxes on Income other than corporation Tax", modify
label def rec_id 23 "Hotel Receipt Tax", modify
label def rec_id 28 "Other Taxes on Income and Expenditure", modify
label def rec_id 29 "Land Revenue", modify
label def rec_id 85 "CGST", modify
label def rec_id 86 "SGST", modify
label values ec_code tax_id
label def tax_id 1 "Txt", modify
label def tax_id 2 "Into", modify
label def tax_id 3 "Mr", modify
label def tax_id 4 "Txo", modify
label def tax_id 5 "Txn", modify
label def tax_id 6 "Pr", modify
label values p_code p_code_id
label def p_code_id 13 "nc", modify

[Nick Cox]

They are similar, but not identical. To see that, try

graph pie rec, over(ec_code) by(year)

The negative values here in rec are troubling, but they disappear on aggregation. Provided that's generally true, pie charts are not ruled out absolutely. They are just not very good at showing the structure of your data. Indeed, your perception that the results are the same is on all fours with the point that very small values or differences are hard to discern on a pie chart, let alone between pie charts.

Edward Tufte said this in 1983:

A table is nearly always better than a dumb pie chart; the only worse design than a pie chart is several of them, for then the viewer is asked to compare quantities located in spatial disarray both within and between pies .... Given their low data-density and failure to order numbers along a visual dimension, pie charts should never be used.

The Visual Display of Quantitative Information. Cheshire, CT: Graphics Press. p.178.

Almost always bar charts out-perform pie charts. Here I used tabplot from the Stata Journal. For other examples, see e.g. https://www.statalist.org/forums/for...updated-on-ssc

* Example generated by -dataex-. To install: ssc install dataex
clear
input int id byte year int head float(ec_code p_code) long rec
 1 22 85 1 13  282984
 1 23 85 1 13  329073
 1 24 85 1 13  369262
 2 22 86 1 13  349992
 2 23 86 1 13  502859
 2 24 86 1 13  510600
 3 22 86 1 13      -9
 3 23 86 1 13       0
 3 24 86 1 13       0
 4 22 86 1 13       0
 4 23 86 1 13       0
 4 24 86 1 13       0
 5 22 86 1 13     -10
 5 23 86 1 13       0
 5 24 86 1 13       0
 6 22 86 1 13       0
 6 23 86 1 13       0
 6 24 86 1 13       0
 7 22 86 1 13  -18613
 7 23 86 1 13       0
 7 24 86 1 13  -22258
 8 22 86 2 13    7332
 8 23 86 2 13       0
 8 24 86 2 13   11519
 9 22 86 3 13     277
 9 23 86 3 13       0
 9 24 86 3 13     400
10 22 86 3 13   19870
10 23 86 3 13       0
10 24 86 3 13   20469
11 22 86 3 13    -152
11 23 86 3 13       0
11 24 86 3 13       0
12 22 86 1 13  465664
12 23 86 1 13  600000
12 24 86 1 13  720066
13 22 86 1 13 -344000
13 23 86 1 13 -400000
13 24 86 1 13 -484214
14 22 86 1 13   71479
14 23 86 1 13   36775
14 24 86 1 13   80000
15 22 86 2 13      62
15 23 86 2 13       0
15 24 86 2 13       0
16 22 86 1 13   44915
16 23 86 1 13       0
16 24 86 1 13   42184
17 22 86 3 13     529
17 23 86 3 13       0
17 24 86 3 13       0
18 22 20 4 13  298575
18 23 20 4 13  330702
18 24 20 4 13  365509
19 22 21 4 13  293864
19 23 21 4 13  321296
19 24 21 4 13  355590
20 22 23 1 13       9
20 23 23 1 13       9
20 24 23 1 13       8
21 22 28 4 13       2
21 23 28 4 13       0
21 24 28 4 13       2
22 22 29 5 13     736
22 23 29 5 13    2000
22 24 29 5 13    2100
23 22 29 5 13      33
23 23 29 5 13     500
23 24 29 5 13     500
24 22 29 5 13      75
24 23 29 5 13       0
24 24 29 5 13       0
25 22 29 6 13     254
25 23 29 6 13      52
25 24 29 6 13      51
26 22 29 5 13       0
26 23 29 5 13       0
26 24 29 5 13       0
27 22 29 3 13       1
27 23 29 3 13       0
27 24 29 3 13       0
28 22 29 6 13     266
28 23 29 6 13      50
28 24 29 6 13     500
29 22 29 5 13    1491
29 23 29 5 13       0
29 24 29 5 13    1000
30 22 29 5 13      26
30 23 29 5 13      50
30 24 29 5 13      55
31 22 29 3 13     136
31 23 29 3 13     900
31 24 29 3 13     900
32 22 29 3 13       2
32 23 29 3 13       0
32 24 29 3 13       0
33 22 29 6 13       1
33 23 29 6 13       0
33 24 29 6 13       0
34 22 29 5 13      52
end
label values year year_id
label def year_id 22 "AE 2021-22", modify
label def year_id 23 "RE 2022-23", modify
label def year_id 24 "BE 2023-24", modify
label values head rec_id
label def rec_id 20 "Corporation Tax", modify
label def rec_id 21 "Taxes on Income other than corporation Tax", modify
label def rec_id 23 "Hotel Receipt Tax", modify
label def rec_id 28 "Other Taxes on Income and Expenditure", modify
label def rec_id 29 "Land Revenue", modify
label def rec_id 85 "CGST", modify
label def rec_id 86 "SGST", modify
label values ec_code tax_id
label def tax_id 1 "Txt", modify
label def tax_id 2 "Into", modify
label def tax_id 3 "Mr", modify
label def tax_id 4 "Txo", modify
label def tax_id 5 "Txn", modify
label def tax_id 6 "Pr", modify
label values p_code p_code_id
label def p_code_id 13 "nc", modify

graph pie rec, over(ec_code) by(year)

tabplot ec_code year [iw=rec], separate(ec_code) showval ytitle("") xtitle("")

[Nick Cox]

Naturally, there is scope for longer value labels and/or a different order on the vertical axis.

[Maneesh Rana]

Thank you

[Maneesh Rana]

I simply draw bar chart with different year with sort over eco_code. The command use
graph bar (sum) rec if year == 22, over(ec_code, sort(ec_code) descending label(angle(forty_five))) blabel(total)

Its saying variable ec_code not found, why so? where I am going wrong..

[Nick Cox]

Is it eco_code or ec_code?

[Maneesh Rana]

Its ec_code, I am trying to draw chart over ec_code, eco_code written by mistake

[Maneesh Rana]

Its ec_code, eco_code written by mistake

[Maneesh Rana]

Hello Nick, please address my problem

[Nick Cox]

I didn't realise that you still had a problem. #7 and #8 seemed to be agreeing with my guess in #6 that you had typed the variable name incorrectly.

With the data example in #1 the code in #5 gives me a different error message:

failed to setup collapse problem
r(198);

but this works fine:

graph bar (sum) rec if year == 22, over(ec_code, sort(1) descending) blabel(total)

I can't see a case for 45 degree labels.

[Maneesh Rana]

Thank you so much. I was eagerly waiting for its resolution. Thanks. Why it was not founding ec_code when used within sort parenthesis.