Chapter Contents |
Previous |
Next |
Details and Examples |
See PARETO13 in the SAS/QC Sample Library |
This example illustrates methods for controlling the order of rows and columns in a comparative Pareto chart.
The following statements create a data set named FAILURE7:
proc format; value procfmt 1 = 'Process A' 2 = 'Process B' ; value dayfmt 1 = 'Monday' 2 = 'Tuesday' 3 = 'Wednesday' 4 = 'Thursday' 5 = 'Friday' ; data failure7; length cause $16 ; format process procfmt. day dayfmt. ; label cause = 'Cause of Failure' process = 'Cleaning Method' day = 'Day of Manufacture' ; input process day cause $16. counts @@; datalines; 1 1 Contamination 15 1 1 Corrosion 2 1 1 Doping 1 1 1 Metallization 2 1 1 Miscellaneous 3 1 1 Oxide Defect 8 1 1 Silicon Defect 1 1 2 Contamination 16 1 2 Corrosion 3 1 2 Doping 1 1 2 Metallization 3 1 2 Miscellaneous 1 1 2 Oxide Defect 9 1 2 Silicon Defect 2 1 3 Contamination 20 1 3 Corrosion 1 1 3 Doping 1 1 3 Metallization 0 1 3 Miscellaneous 3 1 3 Oxide Defect 7 1 3 Silicon Defect 2 1 4 Contamination 12 1 4 Corrosion 1 1 4 Doping 1 1 4 Metallization 0 1 4 Miscellaneous 0 1 4 Oxide Defect 10 1 4 Silicon Defect 1 1 5 Contamination 23 1 5 Corrosion 1 1 5 Doping 1 1 5 Metallization 0 1 5 Miscellaneous 1 1 5 Oxide Defect 8 1 5 Silicon Defect 2 2 1 Contamination 8 2 1 Corrosion 2 2 1 Doping 1 2 1 Metallization 4 2 1 Miscellaneous 2 2 1 Oxide Defect 10 2 1 Silicon Defect 3 2 2 Contamination 9 2 2 Corrosion 0 2 2 Doping 1 2 2 Metallization 2 2 2 Miscellaneous 4 2 2 Oxide Defect 9 2 2 Silicon Defect 2 2 3 Contamination 4 2 3 Corrosion 1 2 3 Doping 1 2 3 Metallization 0 2 3 Miscellaneous 0 2 3 Oxide Defect 10 2 3 Silicon Defect 1 2 4 Contamination 2 2 4 Corrosion 2 2 4 Doping 1 2 4 Metallization 0 2 4 Miscellaneous 3 2 4 Oxide Defect 7 2 4 Silicon Defect 1 2 5 Contamination 1 2 5 Corrosion 3 2 5 Doping 1 2 5 Metallization 0 2 5 Miscellaneous 1 2 5 Oxide Defect 8 2 5 Silicon Defect 2 ;Note that FAILURE7 is similar to the data set FAILURE4 created in Example 29.2. Here, the classification variables PROCESS and DAY are numeric formatted variables, and the formatted values of DAY are Monday through Friday. In Example 29.2, PROCESS and DAY are character variables, and the values of DAY are March 1 through March 5.
The following statements create a two-way comparative Pareto chart for CAUSE in which the rows represent levels of PROCESS and the columns represent levels of DAY:
title 'Pareto Analysis by Process and Day' ; proc pareto data=failure7; vbar cause / class = ( process day ) freq = counts nrows = 2 ncols = 5 cframe = ligr cbars = vigb last = 'Miscellaneous' scale = count catleglabel = 'Failure Causes:' intertile = 1.0 cframeside = ligr cframetop = ligr nohlabel nocurve nlegend ; run;
The chart is shown in Output 29.6.1. The levels of the classification variables are determined by their formatted values. The default order in which the rows and columns are displayed is determined by the internal values of the classification variables, and, consequently, the columns appear in the order of the days of the week.
If DAY had been defined as a character variable with values Monday through Friday, the columns in Output 29.6.1 would have appeared in alphabetical order.
You can override the default order with the ORDER1= and ORDER2= options.
Output 29.6.1: Controlling Row and Column Order
Chapter Contents |
Previous |
Next |
Top |
Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.