Introduction to Splus

In this lab, we will learn data input/output, splus graphcis and basic data analysis funcitons. The tutorial material written by B. Venables and D. Smith is highly recommended. Splus is an object-oriented programming and statistical analysis language. Data in splus is aggregated into "objects" of different types. The most common objects types are vectors, matrices, data frames and lists. Other object types include time series, arrays, factors, etc. Most work in splus is accomplished by functions, which typically create new objects from existing ones. Splus is case sensitive. Create a directory called _Data to store your objects: For h: drive

mkdir h:\Stat524
mkdir h:\Stat524\_Data

For a: drive

mkdir a:\Stat524
mkdir a:\Stat524\_Data

Now create directories to store ordinary data sets

mkdir a:\Dataset
mkdir h:\Dataset

Create a simple data set called temp.txt using microsoft word, save it to either h:\Dataset or a:\Dataset in text format.

Down load a data set from http://www.stat.purdue.edu/~yuzhu/stat524/Datasets/gala.data. Save it to either h:\Dataset or a:\Dataset. Goto standard software\statistical packages\s-plus 4.0, and clik on it. If there is a popup windown asking for directory setup, jsut clik on okay or cancel. And probablly you need try to open s-plus 4.0 again. Though windows version of splus provides user friendly interface, we will focus on the language itself and primarily use the command windows. Your first job every time you start is to tell splus where to store your objects. In lab, we will use h: drive because it seems to be fast.

>attach("h:\\Stat524\\_Data", pos=1)
>search()

search() gives the splus search path list. "h:\Stat524\_Data" occupies the first position to which splus writes new objects. Input data sets into splus. We have already created two data sets in h:\Dataset, which are temp.txt and gala.data.

>temp_read.table("h:\\Dataset\\temp.txt", header=T, row.names=NULL)
>gala_read.table("h:\\Dataset\\gala.data",header=T)
>temp
>gala

There are two ways to get each variable from a data set. If all the analyses will be about gala

>attach(gala)
>Species
>Area
>search()
>detach(2)
>Species

>gala$Species
>gala$Sp
>gala$Area
>gala$Ar

Figure out what each of these commands do by experimentation.

>summary(gala)
>mean(gala$Species)
>min(gala$Sp)
>range(gala$Sp)
>var(gala$Sp)
>median(gala$Sp)
>stem(gala$Sp)
>gala$Endemics/gala$Species
>ratio_gala$Endemics/gala$Species
>summary(ratio)
>larea_log(gala$Area)
>larea
>var(gala)
>cor(gala[1:7])
>cor(gala[,-(1:2)])

You can copy-paste the output to miscrosoft word incorprated into your reports. Splus has very strong graphical capabilities. To create a plot, open a a graphics window:

>win.graph()
>hist(gala$Sp)

You can save, print, and copy-paste a graph. If there are only one graphics window opened, a new graph will replace the old one. You can open more than one graphics widnows. Some time, you want to have several plots on one page, so you need partition the graphics window first:

>par(mfrow=c(2,2))
>hist(gala$Sp)
>hist(gala$Area)
>hist(gala$Elevation)
>hist(gala$Scruz)

Reture to default format

 
>par(mfrow=c(1,1))

Now see if you can figure out what these commands do:

>boxplot(gala$Sp)
>plot(gala$Sp,gala$En)
>plot(gala$Sp,gala$En,xlab="Species",ylab="Endemics",main="Title here")
>plot(gala$Sp,gala$En,xlab="Species",ylab="Endemics",main="Title here",type
="n")
>text(gala$Sp,gala$En,row.names(gala))
>pairs(gala)
>par(mfrow=c(2,3))
>for(i in 2:7) boxplot(gala[,i])
>plot(gala)

While running splus you can get help about a particular commands. For example, you want to know what the command pairs is, you can type

>help(pairs)

If you do not know the name of the command you want to use, you can type

>help()

and then browse. You are encoraged to explore splus further based on the online materials or the books you bought. Before quitting, it is usually a good idea to clean up your workplace, to keep it from being cluttered up with objects you will not use again. To view the names of the objects in the workspace, type

>objects()

To remove unwanted ohjucts, such as the data temp, simply type

>rm(temp)
>objects()

Now, you can quit splus by closing splus window, or by entering

>q()