Getting started with R
Sujit Sahu
25 June 2018
R is very easy to use. To perform 2+2 just type 2+2 and then hit Enter in the command prompt
2+2## [1] 4To find the mean of a bunch of numbers: Form a vector of numbers by the c() command and then try the mean function as below.
The mean is simply the arithmetic mean: \[ \bar{x} = \frac{1}{n} \sum_{i=1}^nx_i \] where \(n\) is the sample size.
mean(c(38, 100, 64, 43, 63, 59, 107, 52, 86, 77))## [1] 68.9Data can be read in various ways, for example, by issuing the read.csv if its a comma separated value file or a read.table if its a tab delimited data file we would like to read. For this to work we first need to tell R which is our current working directoy. We can use the meus or set the working directory as:
setwd("C:/Users/sks/Dropbox/sks/math1024/rmarkdown") # Change this Now we can read data files as follows. We first read the fast food data set. We typed ffood in the second line to see what’s in the object ffood.
ffood <- read.csv("servicetime.csv", head=T) # csv stands for comma separated value file.
# head =T says head = TRUe, so the first row of the file contains the column names.
ffood## AM PM
## 1 38 45
## 2 100 62
## 3 64 52
## 4 43 72
## 5 63 81
## 6 59 88
## 7 107 64
## 8 52 75
## 9 86 59
## 10 77 70For large data sets, we can use the head or tail command to see the top or bottom part of the data set.
head(ffood) # Gets the head of the daya set ## AM PM
## 1 38 45
## 2 100 62
## 3 64 52
## 4 43 72
## 5 63 81
## 6 59 88tail(ffood) # Gets the tail## AM PM
## 5 63 81
## 6 59 88
## 7 107 64
## 8 52 75
## 9 86 59
## 10 77 70The summary command can be used to produce columnwise summaries.
summary(ffood)## AM PM
## Min. : 38.00 Min. :45.00
## 1st Qu.: 53.75 1st Qu.:59.75
## Median : 63.50 Median :67.00
## Mean : 68.90 Mean :66.80
## 3rd Qu.: 83.75 3rd Qu.:74.25
## Max. :107.00 Max. :88.00To see what all columns are there we issue the names command. We access columns by the $ symbol, ffood$AM only gives the AM column of ffood. We can also use the box [] brackets as shown.
names(ffood) ## Prints the column names of the argument data frame. ## [1] "AM" "PM"ffood$AM # Prints the AM values ## [1] 38 100 64 43 63 59 107 52 86 77ffood[,1] # Gets the first column and all rows. ## [1] 38 100 64 43 63 59 107 52 86 77ffood[1:2, ] ## Gets the first two rows and all columns. ## AM PM
## 1 38 45
## 2 100 62ffood[1, 2] ## Gets the first row second column entry ## [1] 45Can you see what values the commands are getting? x[,1] gets the first column and all rows, x[1:2, ] gets the first two rows but all the columns, x[1,2] gets the first row second column entry of x.
Now we read the computer failure data and calculate the data summaries.
cfail <- scan("compfail.txt")
summary(cfail)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.00 1.00 3.00 3.75 5.00 17.00var(cfail)## [1] 11.43204We can tabulate the data by the table command and then we draw a frequency histogram to investigate shape of the data.
table(cfail)## cfail
## 0 1 2 3 4 5 6 7 8 9 10 11 12 13 17
## 12 16 21 12 11 8 7 2 4 2 3 2 2 1 1hist(cfail)
## The Table command shows that the mode is at 2 failures per week. The histogram shows a very skewed distribution. We can draw the boxplot of the data by issuing the boxplot command.
boxplot(cfail)
The median of the data is the thick black line in the middle, and the first and third quartiles are the sides of the box. The two straightlines at the top and bottom connected by the dashed lines are furthest observations from the median but within 1.5 of the inter-quartile range from the median. Suspected outliers are shown as circles outside the whiskers. The boxplot shows positive skewness to the right as well. Now we go to read and explore the other data sets.
wgain <- read.table("wtgain.txt", head=T)
head(wgain)## student initial final
## 1 1 77.56423 76.20346
## 2 2 49.89512 50.34871
## 3 3 60.78133 61.68851
## 4 4 52.16308 53.97745
## 5 5 68.03880 70.30676
## 6 6 47.17357 48.08075summary(wgain)## student initial final
## Min. : 1.00 Min. :42.64 Min. : 43.54
## 1st Qu.:17.75 1st Qu.:53.86 1st Qu.: 54.32
## Median :34.50 Median :60.78 Median : 60.78
## Mean :34.50 Mean :61.72 Mean : 62.59
## 3rd Qu.:51.25 3rd Qu.:68.04 3rd Qu.: 68.49
## Max. :68.00 Max. :99.79 Max. :101.60summary(wgain$final-wgain$initial)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## -2.2680 0.4536 0.9072 0.8672 1.3608 3.6287We can calculate the difference and analyse.
gain <- wgain$final - wgain$initial
summary(gain)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## -2.2680 0.4536 0.9072 0.8672 1.3608 3.6287hist(gain)
boxplot(gain)
We see a little bit of positive weight average weight gain and the histogram indicates a fairly symmetric distribution of differences. The difference itself spans from -2.26 to 3.63 kilograms.
We now read the age guessing data set.
guess <- read.table("guess.txt", head=TRUE, sep=",")
guess## group P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 mae gsize sex
## 1 1 14 -6 5 19 0 -5 -9 -1 -7 -7 7.3 3 F
## 2 2 8 0 5 0 5 -1 -8 -1 -8 0 3.6 4 F
## 3 3 6 -5 6 3 1 -8 -18 1 -9 -6 6.3 4 M
## 4 4 10 -7 3 3 2 -2 -13 6 -7 -7 6.0 2 X
## 5 5 11 -3 4 2 -1 0 -17 0 -14 3 5.5 2 F
## 6 6 13 -3 3 5 -2 -8 -9 -1 -7 0 5.1 3 F
## 7 7 9 -4 3 0 4 -13 -15 6 -7 5 6.6 4 M
## 8 8 11 0 2 8 3 3 -15 1 -7 0 5.0 4 M
## 9 9 6 -2 2 8 3 -8 -7 -1 1 -2 4.0 4 F
## 10 10 11 2 3 11 1 -8 -14 -2 -1 0 5.3 4 Fsummary(guess)## group P1 P2 P3
## Min. : 1.00 Min. : 6.00 Min. :-7.00 Min. :2.00
## 1st Qu.: 3.25 1st Qu.: 8.25 1st Qu.:-4.75 1st Qu.:3.00
## Median : 5.50 Median :10.50 Median :-3.00 Median :3.00
## Mean : 5.50 Mean : 9.90 Mean :-2.80 Mean :3.60
## 3rd Qu.: 7.75 3rd Qu.:11.00 3rd Qu.:-0.50 3rd Qu.:4.75
## Max. :10.00 Max. :14.00 Max. : 2.00 Max. :6.00
## P4 P5 P6 P7
## Min. : 0.00 Min. :-2.00 Min. :-13.00 Min. :-18.0
## 1st Qu.: 2.25 1st Qu.: 0.25 1st Qu.: -8.00 1st Qu.:-15.0
## Median : 4.00 Median : 1.50 Median : -6.50 Median :-13.5
## Mean : 5.90 Mean : 1.60 Mean : -5.00 Mean :-12.5
## 3rd Qu.: 8.00 3rd Qu.: 3.00 3rd Qu.: -1.25 3rd Qu.: -9.0
## Max. :19.00 Max. : 5.00 Max. : 3.00 Max. : -7.0
## P8 P9 P10 mae
## Min. :-2.0 Min. :-14.00 Min. :-7.0 Min. :3.600
## 1st Qu.:-1.0 1st Qu.: -7.75 1st Qu.:-5.0 1st Qu.:5.025
## Median :-0.5 Median : -7.00 Median : 0.0 Median :5.400
## Mean : 0.8 Mean : -6.60 Mean :-1.4 Mean :5.470
## 3rd Qu.: 1.0 3rd Qu.: -7.00 3rd Qu.: 0.0 3rd Qu.:6.225
## Max. : 6.0 Max. : 1.00 Max. : 5.0 Max. :7.300
## gsize sex
## Min. :2.0 F :5
## 1st Qu.:3.0 F :1
## Median :4.0 M :3
## Mean :3.4 X :1
## 3rd Qu.:4.0
## Max. :4.0Here the mae column is the mean absolute error. How can we re-calculate the column just to check that our data has been read correctly? There are many ways to do it. But the following chunk of code does it.
A <- guess[, c(2:11)] ## This only stores the score columns 2 to 11.
A <- abs(A) ## Calculats the absolute value of each entry.
newmae <- apply(A, 1, FUN=mean) ## Used to get the row means as instructed by the value 1.
# Use ?apply to see what it doe
A ## prints A on screen## P1 P2 P3 P4 P5 P6 P7 P8 P9 P10
## 1 14 6 5 19 0 5 9 1 7 7
## 2 8 0 5 0 5 1 8 1 8 0
## 3 6 5 6 3 1 8 18 1 9 6
## 4 10 7 3 3 2 2 13 6 7 7
## 5 11 3 4 2 1 0 17 0 14 3
## 6 13 3 3 5 2 8 9 1 7 0
## 7 9 4 3 0 4 13 15 6 7 5
## 8 11 0 2 8 3 3 15 1 7 0
## 9 6 2 2 8 3 8 7 1 1 2
## 10 11 2 3 11 1 8 14 2 1 0newmae ## prints newmae on screen ## [1] 7.3 3.6 6.3 6.0 5.5 5.1 6.6 5.0 4.0 5.3newmae - guess$mae # check that all values are zero ## [1] 0 0 0 0 0 0 0 0 0 0