使用ggplot在散点图上绘制回归线

Question

I have a quadratic regression model. I would like to add the model's fitted regression line to a scatter plot. My preference is to use ggplot2. I am able to draw the scatter plot but when I use "stat_smooth()" to specify the formula, I get the following warning and the fitted line is not drawn on the scatter plot.

Warning messages: 1: 'newdata' had 80 rows but variables found have 24 rows 2: Computation failed in stat_smooth() : arguments imply differing number of rows: 80, 24

My code is below. Can someone please guide me what should I do differently so that I can get fitted regression line in a scatter plot using ggplot.

Code:

library(gamair)
library(ggplot2)
data(hubble)

names(hubble)[names(hubble) == "y"] <- c("velocity")
names(hubble)[names(hubble) == "x"] <- c("distance")

hubble$distance.sqr <- hubble$distance^2
model2.formula <- hubble$velocity ~ hubble$distance + 
   hubble$distance.sqr - 1
model2.hbl <- lm(model2.formula, data = hubble)
summary(model2.hbl)

model2.sp <- ggplot(hubble, aes(x = distance, y = velocity)) +
  geom_point() + labs(title = "Scatter Plot between Distance & Velocity", 
  x = "Distance", y = "Velocity")
model2.sp + stat_smooth(method = "lm", formula = hubble$velocity ~ 
  hubble$distance + hubble$distance.sqr - 1)

Answer 1

I think the issue here is how you specify the quadratic formula. For the squared term you could use I(x^2) or poly(x, 2) . For example:

ggplot(hubble, aes(x, y)) + 
  geom_point() + 
  stat_smooth(method = "lm", 
              formula = y ~ x + poly(x, 2) - 1) + 
  labs(x = "Distance", y = "Velocity")

Answer 2

here is a MWE based on "mpg" dataset:

library(ggplot2)

ggplot(mpg, aes(x = hwy, y = displ)) + 
  geom_point(shape = 1) + 
  geom_smooth(method = lm, se = FALSE)