Getting Started Using SAS Statistical Analysis System

1
Getting Started Using SAS(Statistical Analysis
System)

• -Anurag Prasad

2
Getting Started Using SASBasics

• Session 1

3
SAS Family

• Base SAS
• SAS/STAT
• SAS/GRAPH
• SAS/ACCESS
• SAS/OR
• SAS/ETS

4
Connecting to SAS

5
Connecting to SAS (Windows)

• Start Exceed
• Start-Menu gt Run gt telnet falaq (and login)
• Type on the prompt
• export DISPLAYltIP addressgt 0.0
• sas
• ( SAS windows open)
• e.g. export DISPLAY172.28.36.2280.0
• sas

6
User Interface

7
SAS Windows

• Program Editor
• Log
• Output
• Results
• Explorer
• Toolbox

8
Program Editor

• Is like a text editor window
• Edit text files, data files, output files
• Edit and submit (run) SAS programs

9
Log

• Contains
• program statements we submitted
• messages from SAS about the execution of the
  program

10
Output

• Contains the results of SAS procedures
• Each observation in a separate row

11
Results

• Helps in managing output of the program

12
Explorer

• Shows the directories and files managed by the
  SAS system
• The library of temporary data sets

13
Toolbox

• Can be used as a command prompt
• to select SAS windows
• to perform basic tasks e.g. copy, save, undo

14
Programming Basics

15
Layout of SAS programs

• Capitalization
• Indentation
• Comments
• Every SAS statement ends with a semicolon.

16
SAS Data Sets

• Rows are Observations
• Columns are Variables
• Two data types numeric character

17
Two parts of a SAS program

• DATA Step
• Creates a data set
• Reads and modifies a data set
• PROC Step
• Performs analysis and functions
• Produces reports

18
DATA Steps built-in L oop

• DATA Steps execute line by line and observation
  by observation

19
A Sample Program
20
SAS Log

• Contains the program statements plus messages
  from SAS about the program
• After DATA Step, it tells how many variables and
  observations are there
• After PROC Step, it tells which procedure took
  how much time

21
Reading SAS Log
22
Reading Output
23
Getting data into SAS

24
Reading Internal Data

• CARDS statement
• is used to indicate internal data
• must be the last statement of the DATA step

25
Example of Internal Data Input

• DATA uspres
• INPUT pres party number
• CARDS
• Adams F 2
• Lincoln R 16
• PROC PRINT
• RUN

26
Reading External Data

• INFILE statement
• is used to read an external data file
• Syntax INFILE lt file name gt

27
Example of External Data Input

• DATA usprez
• INFILE '/users/math/phd/anuragpr/sasfiles/pres.
  txt
• INPUT pres party number
• PROC PRINT
• RUN

28
Different Types of Input

• List Input
• Column Input
• Formatted Input

29
List Input

• Used when
• The values in the data file are separated by at
  least one blank
• Periods (.) indicate missing data
• No embedded space in the character data
• No values are greater than 8 characters

30
Example data for List Input

• Lucky 2.3 1.9 . 3.0
• Spot 4.6 2.5 3.1 .5
• Tubs 7.1 . . 3.8
• Hop 4.5 3.2 1.9 1.8
• Noisy 3.8 1.3 1.8
• 1.5
• Winner 5.7 . . .

31
Program for List Input

• DATA toads
• INFILE toadjump.txt'
• INPUT toadname weight jump1 jump2 jump3
• PROC PRINT
• RUN

32
Column Input

• Used when
• The data file does not have spaces between all
  values
• There are embedded spaces in the character values
• Missing values are blank
• Variables are always found at the same place in
  the data line

33
Example data for Column Input

• Columbia Peaches 35 67 1 10
• Plains Peanuts 210 2 5
• Gilroy Garlics 151035 12 11
• Sacramento Apples 124 85 15 4

34
Program for Column Input

• DATA sales
• INFILE onions.txt'
• INPUT vteam 1-20 csales 21-24 bsales 25-28
  ourhits 29-31 vhits 32-34
• PROC PRINT
• TITLE 'SAS Data Set Sales'
• RUN

35
End of Session 1

36
Getting Started Using SASAdvanced Techniques

• Session 2

37
Working in Program Editor

• d / dd deletes one or more lines
• c / cc copies one or more lines
• m / mm moves one or more lines
• a after this line
• b before this line
• i / i inserts one or more lines
• r / rr repeats one or more new lines
• r / rr repeats number of times
• tc connects two lines of text
• ts splits line at cursor
• cols displays horizontal line ruler

38
Formatted Input

39
INFORMATS

• Three types of informats
• Character ltinformatgtw.
• Numeric ltinformatgtw.d
• Date ltinformatgtw.
• Note w denotes width and
• d denotes the number of decimal places

40
Character Informats

• CHARw.
• Reads character data (does not trim leading or
  trailing blanks)
• INPUT animal CHAR10.
• my cat gt my cat
• my cat gt my cat

41
Character Informats

• w.
• Reads character data (trims leading blanks)
• INPUT animal 10.
• my cat gt my cat
• my cat gt my cat

42
Numeric Informats

• COMMAw.d
• Removes embedded commas and , converts left
  parenthesis to minus sign
• INPUT income COMMA10.
• 1,000,001 gt 1000001
• (1234) gt -1234

43
Numeric Informats

• w.d
• Reads standard numeric data
• INPUT value 5.1
• 1234 gt 123.4
• -12.3 gt -12.3

44
Date Informats

• DATEw.
• Reads dates in form ddmmmyy or ddmmmyyyy
• INPUT day DATE10.
• 1jan1961 gt 366
• 1 jan 1961 gt 366

45
Date Informats

• MMDDYYw.
• Reads dates in form mmddyy or mmddyyyy
• INPUT MMDDYY8.
• 01-01-61 gt 366
• 01/01/61 gt 366

46
INPUT styles

47
Reading Multiple Lines per Observation

• Miami FL
• 90 75
• 97 65
• Nome AK
• 55 44
• 76 98
• INPUT city state
• / var1 var2
• 3 var3 var4

48
Reading Multiple Observations per Line

• Nome AK 34 23 Miami DC 34 65 Raleigh NC
  76 33
• INPUT city state num1 num2 _at__at_
• _at__at_ at the end stops SAS from going to new line
  for each observation

49
Reading Part of a Data File

• a Abc 74
• a Def 97
• b Ghi 87
• a Jkl 79
• b Mno 78
• INPUT type _at_
• IF type b THEN DELETE
• INPUT name 7-10 marks

50
More about Data Sets

51
Creating Permanent Data Sets

• LIBNAME statement defines a libref which points
  to a directory
• DATA lt libref gt.lt member name gt creates a SAS
  data set in that directory whose name is member
  name
• LIBNAME sasbook mysaslib
• DATA sasbook.distance
• miles 23
• km 1.61 miles
• RUN

52
Reading Permanent Data Sets

• Include a LIBNAME statement in program
• Refer to the data set by its two-level name
• LIBNAME example mysaslib
• PROC PRINT DATA example.distance
• TITLE The data set distance
• RUN

53
Writing Raw Data Files

• LIBNAME survey mysaslib
• DATA _NULL_
• SET survey.soap(read the data set soap)
• FILE newfile.txt (create a new file)
• PUT num1 num2 _at_21 chr1 chr2
• RUN

54
Working with Data

55
SAS Functions - Numeric

• INT(arg)
• LOG(arg)
• LOG10(arg)
• MAX(arg,arg,..)
• MIN(arg,arg,..)
• MEAN(arg,arg,..)
• ROUND(arg, round-off-unit)
• SUM(arg,arg,..)

56
SAS Functions - Character

• Character
• LEFT(arg)
• Length(arg)
• SUBSTR(arg,position,n)
• TRIM(arg)
• - Concatenates two character values
• UPCASE(arg)

57
SAS Functions - Date

• DAY(date)
• MDY(month,day,year)
• MONTH(date)
• QTR(date)
• TODAY()

58
IF-THEN Statement

• IF condition THEN action
• A single IF-THEN can have only one action
• Add DO and END for multiple actions
• IF condition THEN DO
• action 1
• action 2
• END

59
Arrays

• ARRAY name () variable-list
• e.g. ARRAY dir (4) north east south west
• Numbered range list
• ARRAY class (100) group1 group100
• INPUT cat4 cat8
• avg MEAN( OF num1 num10 )

60
Sorting, Printing Summarizing Data

61
Sorting our Data PROC SORT

• PROC SORT
• BY var1 var2
• PROC SORT DATA old OUT new
• BY var3 DESCENDING
• Default sort order is ascending/increasing

62
Printing our Data PROC PRINT

• BY variable-list
• Starts a new section in the output for each new
  value of the BY variables.
• ID variable-list
• SUM variable-list
• VAR variable-list
• Specifies which variables are to be printed and
  the order

63
Summarizing Data PROC MEANS

• PROC MEANS options
• N number of non-missing values
• NMISS number of missing values
• MEAN the mean
• STD the standard deviation
• MIN the minimum value
• MAX the maximum value
• RANGE the range
• SUM the sum
• VAR the variance
• SKEWNESS skewness
• KURTOSIS kurtosis
• CV the coeficient of variation

64
Examining Data PROC FREQ

• DATA ufos
• INFILE ufo.txt PAD LRECL 400
• INPUT num time zone descrip 12-46
• PROC FREQ
• TABLES zone time zone
• TITLE UFO Reports
• RUN

65
Data file ufo.txt

• 85845 AM R white with long tail
• 85776 PM C bright white light
• 85873 PM R white and red flashing lights
• 85879 AM R little green men boomboxes
• 86790 PM C throbbing purple light
• 86823 PM R giant toads

66
Visualizing Data PROC PLOT

• PROC PLOT
• PLOT vertical horizontal ( A, B, C,...)
• PLOT height weight
• PLOT height weight name ( first letter )
• BY age_group ( separate plot for each level )
• PLOT csales action C
• bsales action B / OVERLAY

67
Basic SAS Procedures

68
Examining the Distribution PROC UNIVARIATE

• N number of observations
• Mean arithmetic mean
• Std dev standard deviation
• Skewness skewness
• Kurtosis kurtosis
• Max highest value in the data set
• Min lowest value in the data set
• Median median
• Mode most frequently occuring value

69
PROC UNIVARIATE

• Option PLOT produces stem-leaf, box and normal
  probability plot
• Option NORMAL produces tests of normality
• DATA class
• INFILE scores.txt
• INPUT score _at__at_
• PROC UNIVARIATE PLOT NORMAL
• VAR score
• TITLE
• RUN

70
Data file scores.txt

• 56 78 84 73 90 44 76 87 92 75
• 85 67 90 84 74 64 73 78 69 56
• 87 73 100 54 81 78 69 64 73 65

71

72
(No Transcript)
73
(No Transcript)
74
(No Transcript)
75
Examining Correlation PROC CORR

• DATA class
• INFILE study.txt
• INPUT score study exercise
• PROC CORR (Option SPEARMAN)
• VAR study exercise (appear across the top of
  table)
• WITH score (appear down the side)
• RUN

76
Data file study.txt

• 56 6 2 78 7 4 84 5 5 73 4 2 90 6 4
• 44 2 0 76 5 1 87 6 3 92 6 7 75 8 3
• 85 7 1 67 4 2 90 5 5 84 6 5 74 5 2
• 64 4 1 73 8 5 78 5 2 69 6 1 56 4 1
• 87 8 4 73 8 3 100 5 6 54 8 0 81 5 4
• 78 5 2 69 4 1 64 7 1 73 7 3 65 4 4

77
(No Transcript)
78
Simple Regression Analysis

• PROC REG
• MODEL dependent indpendent
• PLOT vertical horizontal symbol
• Options with PLOT (use name followed by a period)
• P predicted values
• R residuals
• STUDENT studentized residual
• U95 upper bound of a 95 c.i.
  for individual prediction
• L95 lower bound

79
Simple Regression Analysis

• DATA hits
• INFILE baseball.txt
• INPUT ht dist
• PROC REG
• MODEL dist ht
• PLOT dist ht P. ht p /OVERLAY
• RUN

80
Data file baseball.txt

• 50 110 49 135 48 129 53 150 48 124
• 50 143 51 126 45 107 53 146 50 154
• 47 136 52 144 47 124 50 133 50 128
• 50 118 48 135 47 129 45 126 48 118
• 45 121 53 142 46 122 47 119 51 134
• 49 130 46 132 51 144 50 132 50 131

81
(No Transcript)
82
(No Transcript)
83
(No Transcript)
84
Analysis of Variance

• PROC ANOVA
• CLASS variable-list
• MODEL dependent effects
• MEANS efects / options
• For one-way anova, the effect is the
  classification variable
• Options with MEANS
• BON (Bonferronis t tests)
• DUNCAN (multiple range)
• T (pairwise t tests)
• SCHEFFE (multiple comparision)
• TUKEY (studentized range)

85
Analysis of Variance

• DATA soft
• INFILE softball.txt
• INPUT team height _at__at_
• PROC ANOVA
• CLASS team
• MODEL height team
• MEANS team
• RUN

86
Data file softball.txt

• red 55 red 48 red 53 red 47 red 51 red 43
• red 45 red 46 red 55 red 54 red 45 red 52
• blue 46 blue 56 blue 48 blue 47 blue 54 blue 52
• blue 49 blue 51 blue 45 blue 48 blue 55 blue 47
• gray 55 gray 45 gray 47 gray 56 gray 49 gray 53
• gray 48 gray 53 gray 51 gray 52 gray 48 gray 47
• pink 53 pink 53 pink 58 pink 56 pink 50 pink 55
• pink 59 pink 57 pink 49 pink 55 pink 56 pink 57
• gold 53 gold 55 gold 48 gold 45 gold 47 gold 56
• gold 55 gold 46 gold 47 gold 53 gold 51 gold 50

87
(No Transcript)
88
(No Transcript)
89
End of Session 2