myTBdata <- read.table("TB_stats.txt", header=TRUE)P01. Introduction to R, part 2: solutions
A. read in a data file
In order to run this, your computer need to know where “TB_stats.txt” is. You could download it here.
What does the “header=TRUE” option mean?
Answer: R reads in the first row as the column headers (or, names)
B. Have a look at the first few lines
# Now let's investigate the data file
head(myTBdata) Country HIV_neg_TB_mortality HIV_pos_TB_mortality
1 Angola 11000 7200
2 Bangladesh 73000 230
3 Brazil 5500 2200
4 Cambodia 8600 440
5 Central_African_Republic 2200 2700
6 China 35000 2600
Total_TB_mortality HIV_pos_TB_incidence Population
1 93000 28000 25000000
2 362000 630 161000000
3 84000 13000 208000000
4 59000 1400 15600000
5 19000 8600 4900000
6 918000 15000 1380000000How many rows can you see? What is the first row?
Answer: 1 row of column names and 6 rows of data
C. What are the names of the columns?
names(myTBdata)[1] "Country" "HIV_neg_TB_mortality" "HIV_pos_TB_mortality"
[4] "Total_TB_mortality" "HIV_pos_TB_incidence" "Population" Is this what you expected?
Answer: We would expect there to be 6 elements to correspond to the 6 columns. Each column name is stored as a separate element in a vector.
D. How many rows and columns are there in your data ?
dim(myTBdata)[1] 30 6What is the first number telling you? And the second?
Answer: dim(…) gives you a 2 element vector; the first number is the number of rows, the second number is the number of columns.
E. How are your data stored?
attributes(myTBdata)$names
[1] "Country" "HIV_neg_TB_mortality" "HIV_pos_TB_mortality"
[4] "Total_TB_mortality" "HIV_pos_TB_incidence" "Population"
$class
[1] "data.frame"
$row.names
[1] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
[26] 26 27 28 29 30What new piece of information have you learned from the ‘attributes()’ function?
Answer: We now know that our data set is stored as a data.frame.
F. Now take a look at some summary statistics for your data
summary(myTBdata) Country HIV_neg_TB_mortality HIV_pos_TB_mortality
Length:30 Min. : 780 Min. : 40
Class :character 1st Qu.: 3725 1st Qu.: 935
Mode :character Median : 14500 Median : 3300
Mean : 40543 Mean :11269
3rd Qu.: 29250 3rd Qu.:10800
Max. :480000 Max. :73000
Total_TB_mortality HIV_pos_TB_incidence Population
Min. : 12000 Min. : 450 Min. :2.140e+06
1st Qu.: 43250 1st Qu.: 5050 1st Qu.:1.560e+07
Median : 122500 Median : 14000 Median :5.370e+07
Mean : 301600 Mean : 33376 Mean :1.546e+08
3rd Qu.: 305500 3rd Qu.: 37500 3rd Qu.:1.325e+08
Max. :2840000 Max. :258000 Max. :1.380e+09 Let’s extract some information from our data
G. First, Calculate the total number of deaths across all countries.
The following two methods should give you the same answer
total_TB_mortality1 <- sum(myTBdata[,2:3]) # method 1
total_TB_mortality2 <- sum(myTBdata$HIV_pos_TB_mortality + myTBdata$HIV_neg_TB_mortality) # method 2Do you think one method is better than the other?
Answer: While method 1 is certaintly shorter, method 2 is generally better for two reasons: 1) it is easier to read because the column names provide information; 2) it is less liable to cause errors (if you for some reason change the indexing of your columns, method 1 might break).
H. Now let’s check that both methods give the same answer.
We’ll use two ways to check this. First, let’s output both answers
total_TB_mortality1[1] 1554340total_TB_mortality2[1] 1554340Now, let’s ask R to check whether they are both equal
total_TB_mortality1==total_TB_mortality2 [1] TRUE# logical expression which gives TRUE if equal and FALSE if notWhy might you prefer to use the second check (using the logical expression) than the first?
Answer: For checking two numbers are the same, it is easy to output two numbers and manually check. However, for larger datasets or for multiple checks, comparison will be much quicker if they are automated.
I. How different is the TB mortality rate in HIV positive persons in Lesotho compared to Zimbabwe?
First, let’s add “mortality rate” as another column in our data frame
myTBdata$Mortality_Per1000 <- 1000 * (myTBdata$HIV_pos_TB_mortality + myTBdata$HIV_neg_TB_mortality)/myTBdata$PopulationNow subset the dataset to extract the TB mortality rate for both Lesotho and Zimbabwe
Lesotho_mortalityrate <- myTBdata[myTBdata$Country=="Lesotho", "Mortality_Per1000"]
Zimbabwe_mortalityrate <- myTBdata[myTBdata$Country=="Zimbabwe", "Mortality_Per1000"]
Relative_Mortality_Rate <- Lesotho_mortalityrate / Zimbabwe_mortalityrateHow many times higher is the mortality rate for TB in Lesotho as it is in Zimbabwe?
paste("The relative mortality rate is", round(Relative_Mortality_Rate, 2), sep=" ")[1] "The relative mortality rate is 5.47"J. Finally in this section, let’s look at what can go wrong when reading in data files.
In order to complete this section, you would need to download readfileexample_1.txt, readfileexample_2.txt, and readfileexample_3.txt.
(a) There is not an equal number of columns in each of the rows.
readFile_a <- read.table("readfileexample_1.txt", header=TRUE)How do you fix this error? Hint: set missing values in the data file to be ‘Not Assigned’ by adding them as NA in the original file. Try running this line again with the updated file.
(b) The wrong delimiter is used
readFile_b <- read.table("readfileexample_2.txt", header=TRUE) Clinic.NumberOfTests.NumberOfPositiveTests
1 1,10,5
2 2,5,2
3 3,6,1
4 4,NA,NA
5 5,NA,NAIs an error given? Check out ‘readFile_b’ - is it correct?
Answer: No error is given but if you type head(readFile_b), the file has not read properly. The file should be read in as a 5 x 3 data.frame. However, instead it is a 5 x 1
How do you fix this? Ask R for help (?read.table) Which option do you need to specify?
Answer: We must specify the delimiter of the text file
readFile_b <- read.table("readfileexample_2.txt", sep = ",", header=TRUE) Clinic NumberOfTests NumberOfPositiveTests
1 1 10 5
2 2 5 2
3 3 6 1
4 4 NA NA
5 5 NA NAIs there another way of fixing this problem?
Answer: We could use the function read.csv() which automatically uses the
readFile_b <- read.csv("readfileexample_2.txt", header=TRUE) Clinic NumberOfTests NumberOfPositiveTests
1 1 10 5
2 2 5 2
3 3 6 1
4 4 NA NA
5 5 NA NA(c) The names are read in as data rows rather than names
readFile_c <- read.csv("readfileexample_2.txt", header=FALSE) V1 V2 V3
1 Clinic NumberOfTests NumberOfPositiveTests
2 1 10 5
3 2 5 2
4 3 6 1
5 4 <NA> <NA>
6 5 <NA> <NA>Is an error given? Check out ‘readFile_c’ - is it correct?
Answer: No, R has read in the first row of the file as regular data points
Type a new line of code to correct this problem (hint: copy-paste from above and change one of the options)
readFile_c <- read.csv("readfileexample_2.txt", header=TRUE) Clinic NumberOfTests NumberOfPositiveTests
1 1 10 5
2 2 5 2
3 3 6 1
4 4 NA NA
5 5 NA NA(d) One of more of the columns contain different classes
readFile_d <- read.table("readfileexample_3.txt", header=TRUE) Clinic Catchment_1000s Doctors
1 1 1.2 10
2 2 5 12
3 3 0.4 3
4 4 1.1 5
5 5 ten 6Is an error given? Check out ‘readFile_d’ - is it correct?
Answer: It appears to look OK. However one of the values has mistakenly been inputted as ‘ten’ rather than ‘10’. This means that the column has been saved as non-numeric.
How do you fix this issue? Hint: check the ‘class’ of the problem column. Try running this line again with an updated file.
Answer: CHANGE THE 5th row, 2nd column to the number 10 in the file and re-run the code