Three ways to load in data in R-format

For this session, we will use a version of the Norwegian ESS (2014). We start by loading the data and giving it a name we like.

1. You can find the data in my R package: RiPraksis

library(RiPraksis)
data(kap6)

2. Alternatively, you can download the file from my webpage. You can do this automatically from R studio. Here, I have already set my working directory, so it suffices to specify that the destination file should be called kap.rda.

download.file("https://siljehermansen.github.io/teaching/beyond-linear-models/kap6.rda",
              destfile = "kap6.rda")

I can now load the R-file into R using the load() function. The data is an R-object that is stored as an .rda-file (R-object file). It is good practice to call the file the same thing as the object when you store it.

load("kap6.rda")

3. … or you can do the download by hand. Navigate to the link in your browser, click to download. Move the file from your download folder to your working folder (where your working directory is set). From R, you can now open the file by clicking: “File” -> “Open file” etc.

Finally, I copy my R object into another object with a name that is easier to write. My main data is usually kalled df.rda

df <- kap6

Eye-ball the data

I always start a session by checking out my data. I can look at the entire spread sheet:

View(df)

I at least check the dimensions (number of observations and variables) and variable names.

dim(df); names(df)

## [1] 1502   14

##  [1] "Parti"         "Indtaegt"      "Uddannelse"    "Kvinde"       
##  [5] "Innvandrer"    "Skepsis"       "KultSkepsis"   "TagerJobs"    
##  [9] "TagerSkat"     "Belastning"    "Arbejdslos"    "Begraenset"   
## [13] "IngenKontrakt" "Prekaritet"

I can also check the first six observations.

head(df)

…and the last six observations

head(df)

two operations at a time

base-R: We can put parentheses in parentheses. The last one comes first. The object comes last.

sum(is.na(df$Skepsis))

## [1] 5

tidyverse: We can create a pipe: The object first, then the first operation, then the operation we want to perform on the result of this operation.

df$Skepsis %>%
  is.na(.) %>% 
  sum()

## [1] 5

There are no rules for what is best. You decide!

base

I begin by making a new variable.

df$Poor <- NA

df$Poor[df$Indtaegt < 5] = 1
df$Poor[df$Indtaegt >= 5] = 0

Now, I can check the frequency table.

table(df$Poor)

## 
##   0   1 
## 853 474

… and plot it

barplot(table(df$Poor))

tidyverse

In tidyverse, we use the case_when()-function to express the conditioning.

df %>%
  mutate("Poor" = case_when(Indtaegt < 5 ~ 1,
                            Indtaegt >= 5 ~ 0)
         )

If this looks ok, then we can store the outcome.

df <- 
  df %>%
  mutate("Poor" = case_when(Indtaegt < 5 ~ 1,
                            Indtaegt >= 5 ~ 0)
  )

A nice tool for conditional recoding is the if… else… algorithm. It is based on a condition and an instruction: if a statement is true, then do this; if the statement is not true, then do that.

base

In base R this is done by the ifelse() function.

df$Poor <- ifelse(df$Indtaegt < 5, 
                  yes = 1,
                  no = 0)

tidyverse

In tidyverse this is handled by the if_else() function.

df <- 
  df %>%
  mutate(Poor = if_else(Indtaegt < 5, 
                        1,
                        0))

This is a true time-saver. Also, you can nest one statement in the other such that you end up with several conditions and outputs.

df <- 
  df %>%
  mutate(Rich = if_else(Indtaegt < 3, 
                        "Poor",
                        if_else(Indtaegt > 7,
                                "Rich",
                                "Medium")))

Look! I’ve got three categories.

df$Rich %>% table

## .
## Medium   Poor   Rich 
##    669    241    417

Group averages

We can make manipulations on parts of the data. Tidyverse is unsurpassed on this one. My favorite function is group_by(). It groups/indexes the data according to a grouping variable (“Poor”) so that we can do operations within each group. Here I calculate the mean level of scepticism towards immigrants among poor and not poor respondents.

df <- 
  df %>%
  group_by(Poor) %>%
  mutate("scepticism_gr" = mean(Skepsis, na.rm = T))

I now have a new variable in the data set called “scepticism_gr”.

I always check if everything went as planned. That’s where simple descriptive statistics are useful. Here, I make a frequency table.

df$scepticism_gr %>% table

## .
## 4.80219349784567 5.24373795761079 5.29598308668076 
##              853              175              474

Hm… I recoded “Poor” to only two values, but here I’ve got three. Let’s look closer at the data. There are many variables in the data frame, so I select only a few.

df %>% select(Parti, scepticism_gr)

## Adding missing grouping variables: `Poor`

Right… Sometimes, there is a missing observation in my “Poor” variable. R has thus greated a third group of “NA” for me. Good to know…

base

I can remove data points using the indexing.

df_poor <- df[df$Poor == 1,]

I can also remove variables using the same indexing.

df_poor <- df_poor[, -3]

tidyverse

In tidyverse we have, as you are now used to, a function to do the same operations.

I can remove observations using the filter().

df_poor <-
  df %>%
  filter(Poor == 1)

I can also select or remove variables using the select() function. I can select variables by specifying their name, or deselect them by adding a an exclamation mark first.

df_poor <-
  df_poor %>%
  select(!Uddannelse)