The gpl manual (and the graphboard) have ways to do a parallel
coordinate plot.
This an example of using the graphboard for a parallel coordinates
plot. (just for demonstration, obviously z-scores make no substantive
sense here).
new file.
new file.
GET
FILE='C:\Program
Files\SPSSInc\PASWStatistics18\Samples\English\world95.sav'.
DATASET NAME world95 WINDOW=FRONT.
descriptives variables =lifeexpm lifeexpf literacy babymort /save.
aggregate outfile = * /break = region
/lifeexpm lifeexpf literacy babymort = mean( Zlifeexpm Zlifeexpf
Zliteracy Zbabymort).
GGRAPH
/GRAPHDATASET NAME="graphdataset"
VARIABLES=lifeexpm[LEVEL=ratio] babymort[LEVEL=ratio]
literacy[LEVEL=ratio]
region[LEVEL=ordinal] lifeexpf[LEVEL=ratio]
MISSING=LISTWISE REPORTMISSING=NO
/GRAPHSPEC SOURCE=VIZTEMPLATE(NAME="Parallel"[LOCATION=LOCAL]
MAPPING( "all"="lifeexpm"[DATASET="graphdataset"]
"all"="lifeexpf"[DATASET="graphdataset"]
"all"="literacy"[DATASET="graphdataset"]
"all"="babymort"[DATASET="graphdataset"]
"Color"="region"[DATASET="graphdataset"]))
VIZSTYLESHEET="Traditional"[LOCATION=LOCAL]
LABEL="Parallel: babymort-literacy-lifeexpf-lifeexpm"
DEFAULTTEMPLATE=NO.
Has anybody modified one of these to do an old fashioned profile plot
where the same scale scale is used on all axes, e.g., z-scores,
T-scores, percents, percentiles, or Likert? If you have would you
please post the syntax? E.g., Here I would want -3 to 3 with
horizontal reference axes at -3, -2, -1,
0,
1, 2, 3.
The profile plot is typically used on aggregated data, e.g., cluster
centroids, repeated measure profiles, etc. The profile plot uses
different "scales". It is unclear how the ticks are determined.
Art Kendall
Social Research Consultants
=====================
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
Art Kendall
Social Research Consultants
Locked
