1.1 Install R and Rstudio:

R: https://cloud.r-project.org/

RStudio: https://www.rstudio.com/products/rstudio/download/#download

1.2 Get aquainted with RStudio

• Examine all the windows and tabs in RStudio. Later, you can organize RStudio in your own way
• Set the working directory setwd(), you can check it by using getwd() without arguments.
• Create an R script file, File -> New file -> R script

4.1 Selection of observations

female.filter <- PT[ ,"gndr"] == 2

PT.females <-  PT[female.filter, c("impfree", "gndr", "eduyrs", "domicil", "agea")]

# or shorter

PT.females <-  PT[ PT$gndr == 2, c("impfree", "gndr", "eduyrs", "domicil", "agea")]

PT.females <- subset(
        x = PT,
        subset = gndr == 2, # select cases, use double equality sign  `==`
        select = c("impfree", "gndr", "eduyrs", "domicil", "agea") # select variables, `с()` merges several values into a vector 
        )

4.2 Multiple filters

wom60 <- PT[ PT$gndr == 2 & PT$agea >= 60,  ]

5.1 Recode in car

If you don’t have it, install car package first. And then:

library("car")
PT$impfree.reversed <- Recode(
  var = PT$impfree,  # original variable 
    recodes = "1=6; 2=5; 3=4; 4=3; 5=2; 6=1; NA=NA; else=NA", # recoding rules, follow pattern 'old value = new value'
  as.numeric=TRUE
  ) 

Let’s check if the recoding went well with cross-tabulation using table() function:

table(PT$impfree.reversed, PT$impfree)

   
      1   2   3   4   5   6
  1   0   0   0   0   0  10
  2   0   0   0   0  38   0
  3   0   0   0  94   0   0
  4   0   0 367   0   0   0
  5   0 367   0   0   0   0
  6 385   0   0   0   0   0

❗️Note that labels are not automatically recoded. In order to recode labels, use rec() function of sjmisc package.

Now recode domicil variable to a new three-category variable: (1) big city or suburbs, (2) small city, (3) village and farmhouse. Add an argument as.factor with value TRUE to acknowledge that it is a nominal variable.

library(car) # Start the `car` package

PT$big.city <- Recode(
  var = PT$domicil,  # original variable 
  recodes = 
    " c(1,2)  = 'Big city or suburbs';             
      c(3)    = 'Small town';  
      c(4,5)  = 'Rural'
    ",    
  as.factor=TRUE   # signifies that the output variable is categorical nominal
  )

Use table() to check if the desired result achieved.

5.2 Optional: manual recoding

Recoding is possible with base R indexing tools.

PT$big.city <- rep(NA, nrow(PT))  #  rep stands for repeat, nrow counts the number of rows, so that empty value "" is repeated as many times as there rows.

PT$big.city[ PT$domicil == 3] <- "Small town"

PT$big.city[PT$domicil == 1 | PT$domicil == 2] <- "Big city or suburbs"
PT$big.city[PT$domicil == 4 | PT$domicil == 5] <- "Rural"

PT$big.city <- as.factor(PT$big.city)

table(PT$big.city, PT$domicil, useNA="always")

                     
                        1   2   3   4   5 <NA>
  Big city or suburbs 256 200   0   0   0    0
  Rural                 0   0   0 359  28    0
  Small town            0   0 425   0   0    0
  <NA>                  0   0   0   0   0    2

6.1 Cross-tabulation / contingency tables / frequencies

table():

table(PT$impfree.reversed,
      PT$big.city,
      useNA="always"
      )

      
       Big city or suburbs Rural Small town <NA>
  1                      5     2          3    0
  2                      9    15         14    0
  3                     29    40         25    0
  4                    131   126        110    0
  5                    129   100        138    0
  6                    151    99        133    2
  <NA>                   2     5          2    0

Important argument useNA="always", otherwise missing values are not shown at all.

To get proportions/percents prop.table(). Argument margin = NULL (defualt) returns proportion based on the total sum, 1 - proportion by rows, 2 - proportions by columns.

table1<- table(PT$impfree.reversed,
               PT$big.city,
               useNA="always"
               )

prop.table(table1,
           margin=2
           )

      
       Big city or suburbs       Rural  Small town        <NA>
  1            0.010964912 0.005167959 0.007058824 0.000000000
  2            0.019736842 0.038759690 0.032941176 0.000000000
  3            0.063596491 0.103359173 0.058823529 0.000000000
  4            0.287280702 0.325581395 0.258823529 0.000000000
  5            0.282894737 0.258397933 0.324705882 0.000000000
  6            0.331140351 0.255813953 0.312941176 1.000000000
  <NA>         0.004385965 0.012919897 0.004705882 0.000000000

At the next step, the numbers can be rounded

p.table <- prop.table(table1, margin=2)

round(p.table, 2)

      
       Big city or suburbs Rural Small town <NA>
  1                   0.01  0.01       0.01 0.00
  2                   0.02  0.04       0.03 0.00
  3                   0.06  0.10       0.06 0.00
  4                   0.29  0.33       0.26 0.00
  5                   0.28  0.26       0.32 0.00
  6                   0.33  0.26       0.31 1.00
  <NA>                0.00  0.01       0.00 0.00

6.2 Central tendency measures

Using mean() and sd(), compute means and standard deviations of variable impfree.

mean(PT$agea, na.rm=TRUE)

[1] 52.04882

mean(PT$impfree.reversed, na.rm=TRUE)

[1] 4.743061

sd(PT$agea, na.rm=TRUE)

[1] 18.30046

sd(PT$impfree.reversed, na.rm=TRUE)

[1] 1.107962

Now try to compute mean and standard deviation using functions sum(), length(), and var().

Correct codes are here

6.3 Group statistics and aggregation

When one variable is continuous and the other categorical, computing group means is meaningful.

Two of the many function for this: by() and aggregate().

aggregate(x       = PT$impfree.reversed, # continuous variable to be aggregated
          by      = list(PT$gndr),             # grouping variable
          FUN     = mean,                # aggregation function
          na.rm   = TRUE                 # remove NAs?
)

There can be more than one grouping variable:

aggregate(PT$impfree.reversed, 
          by=PT[, c("big.city", "gndr")],   
          mean,                      
          na.rm=TRUE                 
          )

Now we can compute standard deviations by replacing mean with sd:

aggregate(PT$impfree.reversed, 
          by=PT[, c("big.city", "gndr")],   
          sd,
          na.rm = TRUE
          )

7.1 Histogram

hist(PT$agea)

7.2 Two continuous variables

plot(PT$eduyrs ~ PT$agea)
abline(v=mean(PT$agea, na.rm=T), col="red")

7.3 Categorical and continuous variable

boxplot(as.vector(PT$eduyrs) ~ PT$big.city)

7.4 Math functions

curve(2*x + 1/sqrt(x), col="red")

8.1 Create an MS Word document

Directly generate a new MS Word document and add table:

library(officer)

cross.tab <- prop.table(table1,margin=2)

table.for.export <- as.data.frame.matrix(cross.tab)

d <- read_docx() #generate an empty Word document

d.appended <- body_add_table(d,                       # the empty doc object
                             table.for.export,        # data.frame to add
                             style = "table_template" # default style
                             )
d.appended <- body_add_par(d.appended,
                           "This is my table and some comments.",
                           style = "Normal"
                           ) 

print(d.appended, target = "data/mydocwithtable.docx") # write the file

[1] "/Users/maksimrudnev/Dropbox/DOCs/кафедра/Анализ  и визуализация в R/2019_ISCTE_Workshop/data/mydocwithtable.docx"

The generated file is here

8.2 Generate html file

It is a more universal way. You can open html documents in browser and then copy to MS Word.

Generate table -> export to html or LaTeX -> copy to MS Word.

library(stargazer)

stargazer(table.for.export, 
          type = "html", 
          out  ="data/table.html", 
          summary = FALSE
          )

<table style="text-align:center"><tr><td colspan="5" style="border-bottom: 1px solid black"></td></tr><tr><td style="text-align:left"></td><td>Big city or suburbs</td><td>Rural</td><td>Small town</td><td>NA</td></tr>
<tr><td colspan="5" style="border-bottom: 1px solid black"></td></tr><tr><td style="text-align:left">1</td><td>0.011</td><td>0.005</td><td>0.007</td><td>0</td></tr>
<tr><td style="text-align:left">2</td><td>0.020</td><td>0.039</td><td>0.033</td><td>0</td></tr>
<tr><td style="text-align:left">3</td><td>0.064</td><td>0.103</td><td>0.059</td><td>0</td></tr>
<tr><td style="text-align:left">4</td><td>0.287</td><td>0.326</td><td>0.259</td><td>0</td></tr>
<tr><td style="text-align:left">5</td><td>0.283</td><td>0.258</td><td>0.325</td><td>0</td></tr>
<tr><td style="text-align:left">6</td><td>0.331</td><td>0.256</td><td>0.313</td><td>1</td></tr>
<tr><td style="text-align:left"></td><td>0.004</td><td>0.013</td><td>0.005</td><td>0</td></tr>
<tr><td colspan="5" style="border-bottom: 1px solid black"></td></tr></table>

Resulting file is here: table.html

8.3 Generate Excel files

write.csv(table.for.export, file="data/exported.table.csv")

Resulting file is here: exported.table.csv

9.1 Solution 1

#1-2
4 + 400

[1] 404

#3
myVector <- c(1,2,3)

#4
myVector10 <- myVector * 10
myVector10

[1] 10 20 30

#5
myVector[2]*10

[1] 20

back

9.2 Solution 2

#1-2 Manual, or automated:

download.file(url = "https://maksimrudnev.com/learn/iscte2019/data/ESS8PT.sav", destfile = "ESS8PT.sav")

#3

install.packages("haven")
library("haven")
PT <- read_sav("ESS8PT.sav")

# Alternative - one-line solution

PT <- haven::read_sav("https://maksimrudnev.com/learn/iscte2019/data/ESS8PT.sav")


# 4
str(PT)

back

9.3 Solution 3

sum(PT$agea, na.rm=TRUE)/length(PT$agea)

[1] 52.04882

sum(PT$impfree.reversed, na.rm=TRUE)/length(PT$impfree.reversed)

[1] 4.709449

sqrt(var(PT$agea, na.rm=TRUE))

[1] 18.30046

sqrt(var(PT$impfree.reversed, na.rm=TRUE))

[1] 1.107962

back