If you have a large dataset that includes a categorical variable, you can use the facets command to produce multiple graphs: one for each level of the categorical variable. In the following example, we will create a graph for each level of ETH (1, 2, and 3) using facets = ETH ~ . Note the tilde sign followed by the period. Enter this syntax:

qplot(HEIGHT, WEIGHT_1, data = T,  geom = "point",  main = "HEIGHT VS. WEIGHT BY ETHNICITY", xlab = "WEIGHT BEFORE TREATMENT (Kg)", ylab = "HEIGHT (cm)", facets = ETH ~ .)

Our facet plot is as follows:

Indeed, we have three scatterplots arranged vertically: one for each level of ETHNICITY. We can also create facet plots across the levels of two factor variables. In the following code, we will create a scatterplot of HEIGHT against WEIGHT_1 for each combination of SMOKE and EXERCISE in each facet, where the two levels of gender are represented by shape and color. Here is the required command:

qplot(HEIGHT, WEIGHT_1, data=T, shape=factor(GENDER), color=factor(GENDER),facets=SMOKE ~ EXERCISE, size=I(3), xlab="HEIGHT", ylab="WEIGHT BEFORE TREATMENT", main = "HEIGHT vs. WEIGHT")

You will get the following output:

Since SMOKE and EXERCISE are both categorical variables of two levels, we ended up with four graphs—one for each combination of the two levels. Note the syntax facets=SMOKE ~ EXERCISE.