Now we have to prepare some more complicated data – for each of the **four wealth categories**, we have to calculate the **percentages** of levels of alcohol consumption. To do that, we first calculate how often each combination of these variables occurs in the dataset, using `count()`

function, but this time with two arguments, one for each variable:

tab <- count(dataset, variable_x, variable_y)

where `variable_x`

and `variable_y`

are the variables we want to analyze.

Then we do something very similar to what we did for the first variable – we calculate the **percentage** of the whole for each combination. Again, we use the `mutate`

function. However, this time we have two variables. We must consider which one is the **main variable** and which is the **grouping variable**. Earlier, we set **wealth category** on the **x-axis**; this makes it the **grouping variable** because it will determine how the data points are grouped.

To set a variable as a grouping variable, we use the `group_by`

function:

tab <- group_by(tab, variable_x)

and then we use our

`mutate()`

function to calculate percentages for each group determined by

`variable_x`

:

tab <- mutate(tab, percent = n / sum(n) * 100)