ggplot with multiple regression lines to show random effects

问题

I am aware of this and this posts. However, I don't seem to get the expected result when I try the following:

The data can be loaded directly from here. The idea is that in a completely made-up data set, the levels of glucose in blood for several athletes at the completion of different races would depend on some fictitious amino acid (AAA):

The call for the plot was:

ggplot(df, aes(x = AAA, y = glucose, color=athletes)) +  
  geom_point() + geom_smooth(method="lm", fill=NA)

And I expected to get different lines for each one of the athletes, instead of one single regression line. My idea was to get something similar to this.

回答1:

something like this?

ggplot(df, aes(x = AAA, y = glucose, color=athletes, group=athletes)) +  
  geom_point() + geom_smooth(method="lm", fill=NA)

or maybe you would prefer this

ggplot(df, aes(x = AAA, y = glucose, color=as.factor(athletes), group=as.factor(athletes))) +  
  geom_point() + geom_smooth(method="lm", fill=NA)

来源：https://stackoverflow.com/questions/34454843/ggplot-with-multiple-regression-lines-to-show-random-effects