TA feedback for hw03

[wvictor14]

Hi Jingyiran,

Great work!

I just wanted to expand on a note on why you lost marks for writing "log(GDPpercap)" in the titles of your plots. I only took -1 off even though it's present in both plots, because I think it is an easy mistake to make.

"log(GDPpercap)" in your axis title suggests that the axis labels are in logged values. For example, $1000 = log(GDPpercap), when this isn't the case. The scale itself is logged, but the original non-logged values are the axis tick labels. Basically, a more accurate (correct) title would be just "GDP per capita ....", and then I would add in text somewhere a note that scale is logged.

most data points are concentrated at the lower 75% quantile for Africa,

By definition of a quantile, and boxplots, exactly 75% of all data points are below the 75% quantile for every continent.

Your ggplot2 code has some redundancy:

# too verbose
ggplot(aes(x=year, y=mean.lifeexp, group=continent)) +
  geom_line(aes(color=continent))+
  geom_point(aes(color=continent))

# aesthetics transfer from ggplot layer to other layers
ggplot(aes(x=year, y=mean.lifeexp, color=continent)) +
  geom_line()+
  geom_point()

[lijingyiran]

Thanks so much for pointing out the errors, that's very helpful!!! I keep those in mind for future assignments.