Visualizing Categorical Data: sieveplot

The sieveplot macro ( get sieveplot.sas)

Macro program for sieve diagrams

The SIEVEPLOT macro provides a more convenient macro interface to the SAS/IML modules for sieve diagrams, so that you can input a SAS data set (in frequency form), rather than constructing the arrays used by the IML modules. It also provides the means to produce multiple sieve plots for two-way tables stratified by other BY= variables.

Method

Usage

The arguments may be listed within parentheses in any order, separated by commas. For example:

  %sieveplot(data=haireye, var=Eye Hair, filltype=obsp)

Parameters

DATA=

Name of input dataset [Default: DATA=_LAST_]

VAR=

Names of factor variables, including BY= variables. The first two variables are used as the rows and columns of the sieve diagram.

BY=

Specifies the names of one (or more) By variables. Partial sieve plots are produced for each combination of the levels of the BY= variables. The BY= variable(s) *must* be listed among the VAR= variables.

COUNT=

Name of the frequency variable [Default: COUNT=COUNT]

COLORS=

Colors for + and - residuals [Default: COLORS=BLUE RED]

FILLTYPE=

Specifies how each cell is filled, one of OBS, OBSP, DEV, EXL, EXP [Default: FILLTYPE=OBS] OBS: fill cells in proportion to observed frequencies 'OBSP: like OBS, but also write obs. freq. in cell 'DEV: fill in proportion to deviations from independence. 'EXL: no fill, write expected frequency in cell 'EXP: expfill, write expected frequency in cell

MARGINS=

Empty or TOTAL

HTEXT=

Height of text labels [Default: HTEXT=1.5]

FONT=

Font for text labels

TITLE=

Title for plot(s)

GOUT=

The name of the graphics catalog

NAME=

The name of the graph in the graphic catalog

Example

%include macros(sieveplot);        *-- or include in an autocall library;

  *-- Three way table, Hair color x Eye color x Sex;
  %include catdata(hairdat3);

  *-- Separate plots for males and females;
  %sieveplot(data=haireye, var=Eye Hair Sex, by=Sex, title=Hair-Eye data--,
      filltype=obsp);

  *-- Collapse the table over Sex;
  %table(data=haireye, var=Eye Hair, order=data,
       weight=count, out=haireye2);

  %sieveplot(data=haireye2, var=Eye Hair, filltype=obsp);

See also