SAS job

Data Management and Business Intelligence
5CC519

www.derby.ac.uk/engtech

Lecture 5 – Picturing Distributions and Outliers in SAS

Dr John Panneerselvam

College of Engineering and Technology, University of Derby

*

Distributions

*

Descriptive Analytics

The goals when you are describing data are to

• screen for unusual data values
• inspect the spread and shape of continuous variables
• characterize the central tendency
• draw preliminary conclusions about your data.

*

A Normal Distribution

*

The UNIVARIATE Procedure

General form of the UNIVARIATE procedure:

PROC UNIVARIATE DATA=SAS-data-set <options>;
VAR variables;

CDFPLOT <variables> < / options> ;

HISTOGRAM variables </ options>;

PROBPLOT variables </ options>;

QQPLOT <variables> < / options> ;

VAR variables ;

WEIGHT variable ;

RUN;

*

Descriptive Statistics

• Mean
• Sum
• Sum of the Weights
• Standard Deviation
• Variance
• Covariance
• Skewness
• Kurtosis

*

Creating Histogram

title 'Analysis of Plating Thickness';

proc univariate data=Trans noprint;

histogram Thick;

run;

*

Two Way Comparative Histogram

title 'Results of Supplier Training Program';

proc univariate data=Disk noprint;

class Supplier Year;

histogram Width / intertile = 1.0

vaxis = 0 10 20 30

ncols = 2

nrows = 2;

run;

*

Histogram with a Fit

proc univariate data=Robots;

histogram Length /

beta(theta=10 scale=0.5 fill)

href = 10

hreflabel = 'Lower Bound'

odstitle = 'Fitted Beta

Distribution of Offsets';

inset n = 'Sample Size' /

pos=ne cfill=blank;

run;

*

The UNIVARIATE Procedure

*

CDF with a Fit

proc univariate data=Cord noprint;

cdf Strength / normal;

inset normal(mu sigma);

run;

*

Quantile Fit

symbol v=plus;

title 'Two-Parameter Lognormal Q-Q Plot for Diameters';

ods graphics off;

proc univariate data=ModifiedMeasures noprint;

qqplot LogDiameter / normal(mu=est sigma=est)

square

vaxis=axis1;

inset n mean (5.3) std (5.3) / pos = nw header = 'Summary Statistics';

axis1 label=(a=90 r=0);

run;

*

Outliers

*

Box-and-Whisker Plots

The mean is denoted by a ◊.

largest point <= 1.5 I.Q. from the box

the 75th percentile

the 25th percentile

the 50th percentile (median)

smallest point <= 1.5 I.Q. from the box

outliers > 1.5 I.Q. from the box

1.5 I.Q.

*

The BOXPLOT Procedure

General form of the UNIVARIATE procedure:

PROC BOXPLOT options ;

BY variables ;

INSET keywords </options> ;

INSETGROUP keywords </ options> ;

PLOT analysis-variable*group-variable <(block-variables)> <=symbol-variable> </ options > ;

Run;

*

Creating a Box Plot

proc boxplot data=Turbine;

plot KWatts*Day / boxstyle = schematic

outbox = OilSchematic;

run;

*

Enhancing a Box Plot

title 'Box Plot for Power Output';

proc boxplot data=Turbine;

plot KWatts*Day;

inset min mean max stddev / header = 'Overall Statistics‘ pos = tm;

insetgroup min max /

header = 'Extremes by Day';

run;

*

• Big Data Fundamentals Concepts, Drivers and Techniques (Available in Week 5)
• SAS Programming 1: Essentials- SAS Institute.
• SAS: The Power to Know- Official documentation.

www.derby.ac.uk/engtech

References

*

www.derby.ac.uk/engtech

Data Management and Business Intelligence 5CC519

www.derby.ac.uk/engtech

Software Testing

Dr John Panneerselvam

College of Engineering and Technology, University of Derby

Sensitivity: Internal