How do I modify a stacked bar chart to give it a pyramid-like look?

[James Phillips-5]

I would like to modify a stacked bar chart to present some survey data so that the percent of positive responses is on the right side of the origin and the percent of negative or neutral responses is on the left side. Both sides will originate from the center at 0%. Below is my example data and a starting point for my GPL code. I am new to GPL and am lost on how to do this. Any ideas on how I can split the graph?

Thank you for your help,

Jim

-------------


DATA LIST LIST/
customer_id question_number responses (F4, F2, A30).
BEGIN DATA
1234 1 '4 - agree'
4567 1 '3 - neutral'
7891 1 '2 - disagree'
2345 1 '3 - neutral'
6751 1 '2 - disagree'
9521 1 '3 - neutral'
7415 1 '2 - disagree'
2651 1 '5 - strongly agree'
4152 1 '3 - neutral'
3655 1 '2 - disagree'
2587 1 '3 - neutral'
4125 1 '4 - agree'
3658 1 '3 - neutral'
4157 1 '1 - strongly disagree'
2541 1 '3 - neutral'
8521 1 '4 - agree'
9514 1 '4 - agree'
4125 1 '4 - agree'
3652 1 '3 - neutral'
8569 1 '1 - strongly disagree'
1234 2 '3 - neutral'
4567 2 '2 - disagree'
7891 2 '3 - neutral'
2345 2 '4 - agree'
6751 2 '3 - neutral'
9521 2 '1 - strongly disagree'
7415 2 '3 - neutral'
2651 2 '4 - agree'
4152 2 '5 - strongly agree'
3655 2 '4 - agree'
2587 2 '3 - neutral'
4125 2 '2 - disagree'
3658 2 '2 - disagree'
4157 2 '3 - neutral'
2541 2 '2 - disagree'
8521 2 '3 - neutral'
9514 2 '2 - disagree'
4125 2 '3 - neutral'
3652 2 '2 - disagree'
8569 2 '3 - neutral'
1234 3 '4 - agree'
4567 3 '3 - neutral'
7891 3 '2 - disagree'
2345 3 '3 - neutral'
6751 3 '4 - agree'
9521 3 '3 - neutral'
7415 3 '2 - disagree'
2651 3 '3 - neutral'
4152 3 '5 - strongly agree'
3655 3 '4 - agree'
2587 3 '3 - neutral'
4125 3 '4 - agree'
3658 3 '5 - strongly agree'
4157 3 '3 - neutral'
2541 3 '4 - agree'
8521 3 '5 - strongly agree'
9514 3 '2 - disagree'
4125 3 '1 - strongly disagree'
3652 3 '2 - disagree'
8569 3 '3 - neutral'
1234 4 '3 - neutral'
4567 4 '2 - disagree'
7891 4 '3 - neutral'
2345 4 '4 - agree'
6751 4 '3 - neutral'
9521 4 '1 - strongly disagree'
7415 4 '3 - neutral'
2651 4 '4 - agree'
4152 4 '4 - agree'
3655 4 '4 - agree'
2587 4 '3 - neutral'
4125 4 '1 - strongly disagree'
3658 4 '3 - neutral'
4157 4 '2 - disagree'
2541 4 '5 - strongly agree'
8521 4 '3 - neutral'
9514 4 '2 - disagree'
4125 4 '3 - neutral'
3652 4 '4 - agree'
8569 4 '3 - neutral'
END DATA.


* Chart Builder.
GGRAPH
/GRAPHDATASET NAME="graphdataset" VARIABLES=question_number COUNT()[name="COUNT"] responses MISSING=VARIABLEWISE REPORTMISSING=NO
/GRAPHSPEC SOURCE=INLINE.
BEGIN GPL
SOURCE: s=userSource(id("graphdataset"))
DATA: question_number=col(source(s), name("question_number"), unit.category())
DATA: COUNT=col(source(s), name("COUNT"))
DATA: responses=col(source(s), name("responses"), unit.category())
COORD: rect(dim(1,2), transpose())
GUIDE: axis(dim(1), label("Question Number"))
GUIDE: axis(dim(2), label("Percent"))
GUIDE: legend(aesthetic(aesthetic.color.interior), label("Responses"))
SCALE: cat(dim(1), reverse())
SCALE: linear(dim(2), include(0))
SCALE: cat(aesthetic(aesthetic.color.interior), include("1 - strongly disagree", "2 - disagree", "3 - neutral", "4 - agree", "5 - strongly agree"), reverse(), aestheticMissing(color.black))
ELEMENT: interval.stack(position(summary.percent(question_number*COUNT, base.coordinate(dim(1)))), color.interior(responses), shape.interior(shape.square))
END GPL.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[ViAnn Beadle]

Think of this as a population pyramid chart. You're going to have to recode
provide a splitter variable by recoding your response variable into a binary
variable with positive values set to one category and negative to another
and neutral set to missing. Give that you've got a long string variable here
to represent your value, it's simpler to just use AUTORECODE to get a
variable with values 1 to 5 and then recode from there. The following
replaces your GGRAPH code:

AUTORECODE VARIABLES=responses
  /INTO respval
  /PRINT.

recode respval (1, 2=1)(3,4=2) into resplit.
missing values resplit (3).
value labels resplit 1 "Negative" 2 "Positive".
variable labels resplit "Dichtomized Response".



* Chart Builder.
GGRAPH
  /GRAPHDATASET NAME="graphdataset" VARIABLES=COUNT()[name="COUNT"]
question_number
    resplit[LEVEL=NOMINAL] MISSING=LISTWISE REPORTMISSING=NO
  /GRAPHSPEC SOURCE=INLINE.
BEGIN GPL
  SOURCE: s=userSource(id("graphdataset"))
  DATA: COUNT=col(source(s), name("COUNT"))
  DATA: question_number=col(source(s), name("question_number"),
unit.category())
  DATA: resplit=col(source(s), name("resplit"), unit.category())
  COORD: transpose(mirror(rect(dim(1,2))))
  GUIDE: axis(dim(1), opposite(), label("Question Number"))
  GUIDE: axis(dim(2), label("Percent"))
  GUIDE: axis(dim(3), label("Response"), opposite(), gap(0px))
  GUIDE: legend(aesthetic(aesthetic.color), null())
  SCALE: cat(dim(1), reverse())
  SCALE: linear(dim(2), include(0))
  SCALE: cat(dim(3), include("1.00", "2.00"))
  ELEMENT: interval(position(summary.percent(question_number*COUNT*resplit,
base.all(acrossPanels()))),
      color.interior(resplit), label(COUNT))
END GPL.

Notes:
Note the use of the mirror function to create mirrored values on the COORD
statement.
A population pyramid uses panels rather than clustering. This requires a
different percentage base to get the right percents.
I've added a label of the actual counts to the bars because relying upon
percents with that middle category out might be misleading.
Chart Builder generates population pyramids but doesn't provide the
percentage option.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
James Phillips
Sent: Wednesday, December 03, 2008 11:14 AM
To: [hidden email]
Subject: How do I modify a stacked bar chart to give it a pyramid-like look?

I would like to modify a stacked bar chart to present some survey data so
that the percent of positive responses is on the right side of the origin
and the percent of negative or neutral responses is on the left side. Both
sides will originate from the center at 0%. Below is my example data and a
starting point for my GPL code. I am new to GPL and am lost on how to do
this. Any ideas on how I can split the graph?

Thank you for your help,

Jim

-------------


DATA LIST LIST/
customer_id question_number responses (F4, F2, A30).
BEGIN DATA
1234 1 '4 - agree'
4567 1 '3 - neutral'
7891 1 '2 - disagree'
2345 1 '3 - neutral'
6751 1 '2 - disagree'
9521 1 '3 - neutral'
7415 1 '2 - disagree'
2651 1 '5 - strongly agree'
4152 1 '3 - neutral'
3655 1 '2 - disagree'
2587 1 '3 - neutral'
4125 1 '4 - agree'
3658 1 '3 - neutral'
4157 1 '1 - strongly disagree'
2541 1 '3 - neutral'
8521 1 '4 - agree'
9514 1 '4 - agree'
4125 1 '4 - agree'
3652 1 '3 - neutral'
8569 1 '1 - strongly disagree'
1234 2 '3 - neutral'
4567 2 '2 - disagree'
7891 2 '3 - neutral'
2345 2 '4 - agree'
6751 2 '3 - neutral'
9521 2 '1 - strongly disagree'
7415 2 '3 - neutral'
2651 2 '4 - agree'
4152 2 '5 - strongly agree'
3655 2 '4 - agree'
2587 2 '3 - neutral'
4125 2 '2 - disagree'
3658 2 '2 - disagree'
4157 2 '3 - neutral'
2541 2 '2 - disagree'
8521 2 '3 - neutral'
9514 2 '2 - disagree'
4125 2 '3 - neutral'
3652 2 '2 - disagree'
8569 2 '3 - neutral'
1234 3 '4 - agree'
4567 3 '3 - neutral'
7891 3 '2 - disagree'
2345 3 '3 - neutral'
6751 3 '4 - agree'
9521 3 '3 - neutral'
7415 3 '2 - disagree'
2651 3 '3 - neutral'
4152 3 '5 - strongly agree'
3655 3 '4 - agree'
2587 3 '3 - neutral'
4125 3 '4 - agree'
3658 3 '5 - strongly agree'
4157 3 '3 - neutral'
2541 3 '4 - agree'
8521 3 '5 - strongly agree'
9514 3 '2 - disagree'
4125 3 '1 - strongly disagree'
3652 3 '2 - disagree'
8569 3 '3 - neutral'
1234 4 '3 - neutral'
4567 4 '2 - disagree'
7891 4 '3 - neutral'
2345 4 '4 - agree'
6751 4 '3 - neutral'
9521 4 '1 - strongly disagree'
7415 4 '3 - neutral'
2651 4 '4 - agree'
4152 4 '4 - agree'
3655 4 '4 - agree'
2587 4 '3 - neutral'
4125 4 '1 - strongly disagree'
3658 4 '3 - neutral'
4157 4 '2 - disagree'
2541 4 '5 - strongly agree'
8521 4 '3 - neutral'
9514 4 '2 - disagree'
4125 4 '3 - neutral'
3652 4 '4 - agree'
8569 4 '3 - neutral'
END DATA.


* Chart Builder.
GGRAPH
/GRAPHDATASET NAME="graphdataset" VARIABLES=question_number
COUNT()[name="COUNT"] responses MISSING=VARIABLEWISE REPORTMISSING=NO
/GRAPHSPEC SOURCE=INLINE.
BEGIN GPL
SOURCE: s=userSource(id("graphdataset"))
DATA: question_number=col(source(s), name("question_number"),
unit.category())
DATA: COUNT=col(source(s), name("COUNT"))
DATA: responses=col(source(s), name("responses"), unit.category())
COORD: rect(dim(1,2), transpose())
GUIDE: axis(dim(1), label("Question Number"))
GUIDE: axis(dim(2), label("Percent"))
GUIDE: legend(aesthetic(aesthetic.color.interior), label("Responses"))
SCALE: cat(dim(1), reverse())
SCALE: linear(dim(2), include(0))
SCALE: cat(aesthetic(aesthetic.color.interior), include("1 - strongly
disagree", "2 - disagree", "3 - neutral", "4 - agree", "5 - strongly
agree"), reverse(), aestheticMissing(color.black))
ELEMENT: interval.stack(position(summary.percent(question_number*COUNT,
base.coordinate(dim(1)))), color.interior(responses),
shape.interior(shape.square))
END GPL.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD