How do I create a ggplot in R from a non-linear model using the mgcv package?

Question

I have a non-linear survival model which I have coded using the mgcv package. I can produce a regular plot, but I would like to be able to do code a ggplot2 instead. How do I go about this?

Here is my code:

df <- structure(list(SurvYear =c(3L, 2L, 3L, 6L, 8L, 3L, 5L, 2L, 9L, 
8L, 1L, 7L, 1L, 4L, 6L, 8L, 2L, 5L, 1L, 1L, 7L, 1L, 5L, 3L, 2L, 
1L, 9L, 1L, 5L, 2L, 2L, 1L, 2L, 3L, 4L, 8L, 7L, 2L, 2L, 6L, 9L, 
7L, 3L, 9L, 6L, 8L, 2L, 8L, 2L, 1L, 1L, 6L, 5L, 3L, 3L, 7L, 2L, 
4L, 5L, 2L, 3L, 7L, 4L, 1L, 2L, 2L, 3L, 5L, 1L, 9L, 2L, 2L, 3L, 
9L, 6L, 2L, 2L, 4L, 3L, 1L, 9L, 7L, 3L, 1L, 2L, 1L, 6L, 3L, 1L, 
5L, 6L, 5L, 6L, 4L, 2L, 1L, 3L, 1L, 1L, 3L, 4L, 3L, 8L, 9L, 7L, 
6L, 3L, 5L, 2L, 7L, 9L, 2L, 1L, 1L, 1L, 1L, 2L, 1L, 2L, 9L, 1L, 
4L, 8L, 1L, 8L, 1L, 1L, 8L, 5L, 2L, 9L, 4L, 8L, 4L, 9L, 2L, 2L, 
3L, 2L, 9L, 3L, 2L, 1L, 3L, 2L, 1L, 9L, 9L, 2L, 1L, 1L, 1L, 2L, 
9L, 1L, 5L, 1L, 6L, 9L, 3L, 2L, 2L, 5L, 7L, 4L, 2L, 7L, 2L, 4L, 
5L, 3L, 3L, 9L, 2L, 6L, 1L, 3L, 4L, 5L, 9L, 8L, 1L, 2L, 8L, 2L, 
9L, 1L, 7L, 3L, 3L, 1L, 6L, 3L, 4L, 9L, 1L, 3L, 4L, 4L, 2L, 7L, 
2L, 3L, 1L, 1L, 7L, 2L, 1L, 1L, 2L, 1L, 9L, 1L, 2L, 9L, 1L, 1L, 
2L, 3L, 7L, 3L, 1L, 1L, 2L, 5L, 4L, 6L, 7L, 1L, 9L, 2L, 1L, 8L, 
1L, 2L, 1L, 4L, 2L, 3L, 3L, 9L, 9L, 9L, 4L, 1L, 1L, 4L, 9L, 3L, 
1L, 1L, 3L, 3L, 4L, 1L, 1L, 1L, 1L, 6L, 9L, 1L, 1L, 8L, 1L, 3L, 
3L, 8L, 3L, 5L, 1L, 2L, 1L, 2L, 4L, 3L, 1L, 6L, 1L, 4L, 8L, 1L, 
3L, 2L, 2L, 3L, 6L, 2L, 1L, 1L, 1L, 9L, 3L, 1L, 7L, 3L, 9L, 1L, 
9L, 5L, 4L), Gender = c(1L, 0L, 1L, 1L, 0L, 0L, 0L, 1L, 0L, 0L, 
1L, 0L, 0L, 0L, 1L, 1L, 1L, 0L, 1L, 1L, 0L, 1L, 1L, 1L, 1L, 0L, 
1L, 0L, 1L, 1L, 1L, 0L, 0L, 1L, 1L, 1L, 0L, 0L, 1L, 1L, 1L, 1L, 
0L, 0L, 1L, 0L, 1L, 0L, 1L, 1L, 1L, 0L, 1L, 0L, 0L, 1L, 1L, 0L, 
1L, 0L, 1L, 1L, 1L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 1L, 0L, 0L, 0L, 
1L, 1L, 1L, 1L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 
0L, 0L, 1L, 1L, 1L, 0L, 1L, 1L, 0L, 0L, 1L, 0L, 1L, 1L, 0L, 0L, 
0L, 0L, 1L, 1L, 1L, 0L, 1L, 1L, 1L, 1L, 0L, 1L, 1L, 1L, 0L, 1L, 
0L, 1L, 1L, 1L, 1L, 1L, 1L, 0L, 1L, 1L, 0L, 0L, 0L, 0L, 0L, 1L, 
1L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 0L, 1L, 1L, 1L, 0L, 
0L, 1L, 0L, 0L, 0L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 0L, 1L, 1L, 0L, 
1L, 0L, 1L, 0L, 1L, 1L, 1L, 1L, 0L, 0L, 1L, 0L, 1L, 0L, 1L, 1L, 
1L, 1L, 1L, 0L, 0L, 0L, 0L, 1L, 1L, 1L, 0L, 1L, 0L, 0L, 1L, 0L, 
1L, 1L, 1L, 1L, 1L, 0L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 0L, 1L, 
1L, 0L, 1L, 1L, 0L, 1L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 0L, 1L, 1L, 0L, 1L, 0L, 1L, 0L, 1L, 0L, 1L, 1L, 1L, 
1L, 1L, 0L, 0L, 0L, 1L, 0L, 1L, 0L, 1L, 1L, 1L, 1L, 0L, 1L, 0L, 
0L, 0L, 0L, 0L, 0L, 0L, 0L, 1L, 0L, 1L, 0L, 1L, 1L, 0L, 1L, 1L, 
0L, 0L, 1L, 1L, 0L, 1L, 1L, 1L, 0L, 0L, 0L, 1L, 1L, 0L, 1L, 0L, 
1L, 1L), Age = c(63L, 66L, 34L, 43L, 63L, 21L, 24L, 44L, 52L, 
59L, 27L, 32L, 30L, 20L, 56L, 55L, 35L, 26L, 53L, 43L, 39L, 19L, 
34L, 28L, 19L, 24L, 50L, 22L, 58L, 24L, 50L, 25L, 37L, 30L, 51L, 
69L, 23L, 49L, 22L, 46L, 58L, 31L, 23L, 53L, 59L, 25L, 38L, 44L, 
34L, 49L, 19L, 39L, 24L, 51L, 29L, 27L, 48L, 77L, 22L, 43L, 59L, 
49L, 60L, 51L, 49L, 47L, 50L, 44L, 41L, 44L, 50L, 42L, 46L, 54L, 
35L, 21L, 26L, 26L, 40L, 21L, 48L, 49L, 20L, 20L, 32L, 37L, 22L, 
36L, 46L, 28L, 39L, 35L, 51L, 39L, 49L, 57L, 46L, 18L, 52L, 47L, 
27L, 32L, 23L, 43L, 42L, 57L, 22L, 40L, 19L, 58L, 71L, 55L, 42L, 
20L, 51L, 21L, 20L, 61L, 36L, 54L, 19L, 35L, 38L, 41L, 34L, 22L, 
41L, 42L, 56L, 50L, 53L, 53L, 48L, 22L, 59L, 27L, 28L, 32L, 37L, 
68L, 24L, 26L, 61L, 21L, 20L, 20L, 50L, 62L, 61L, 29L, 18L, 40L, 
67L, 43L, 25L, 43L, 22L, 56L, 47L, 41L, 40L, 43L, 27L, 37L, 61L, 
35L, 23L, 54L, 38L, 38L, 39L, 45L, 49L, 63L, 49L, 44L, 44L, 23L, 
37L, 58L, 61L, 25L, 18L, 59L, 25L, 51L, 40L, 27L, 42L, 22L, 38L, 
22L, 45L, 33L, 32L, 36L, 53L, 52L, 19L, 45L, 53L, 27L, 65L, 25L, 
53L, 57L, 29L, 23L, 62L, 36L, 56L, 59L, 41L, 61L, 44L, 24L, 21L, 
38L, 29L, 55L, 33L, 18L, 21L, 19L, 65L, 24L, 59L, 34L, 25L, 45L, 
48L, 18L, 41L, 61L, 32L, 37L, 21L, 20L, 57L, 25L, 65L, 50L, 61L, 
32L, 27L, 19L, 50L, 63L, 19L, 45L, 20L, 36L, 20L, 19L, 53L, 39L, 
50L, 20L, 24L, 57L, 28L, 21L, 39L, 49L, 21L, 20L, 39L, 20L, 44L, 
19L, 39L, 53L, 29L, 60L, 43L, 21L, 23L, 30L, 42L, 42L, 51L, 35L, 
50L, 51L, 56L, 52L, 22L, 36L, 56L, 28L, 57L, 20L, 47L, 48L, 65L, 
71L, 21L, 70L, 23L, 63L), Highest_Educationmx = c(4L, 5L, 3L, 
2L, 3L, 2L, 3L, 1L, 3L, 1L, 7L, 3L, 2L, 3L, 3L, 2L, 6L, 2L, 3L, 
6L, 3L, 2L, 2L, 7L, 2L, 1L, 2L, 3L, 6L, 3L, 5L, 3L, 5L, 6L, 2L, 
1L, 5L, 2L, 5L, 1L, 1L, 3L, 2L, 3L, 1L, 7L, 5L, 4L, 7L, 3L, 1L, 
1L, 6L, 3L, 3L, 2L, 4L, 6L, 5L, 4L, 2L, 6L, 1L, 3L, 4L, 2L, 1L, 
5L, 5L, 3L, 1L, 5L, 3L, 3L, 1L, 4L, 2L, 3L, 5L, 3L, 1L, 4L, 2L, 
1L, 2L, 7L, 2L, 5L, 3L, 2L, 6L, 1L, 1L, 3L, 4L, 1L, 5L, 1L, 3L, 
4L, 2L, 7L, 2L, 4L, 4L, 7L, 4L, 6L, 3L, 1L, 2L, 1L, 5L, 5L, 1L, 
5L, 2L, 7L, 3L, 4L, 2L, 4L, 2L, 4L, 2L, 2L, 4L, 1L, 2L, 1L, 2L, 
6L, 1L, 2L, 5L, 2L, 2L, 5L, 1L, 6L, 5L, 2L, 1L, 2L, 1L, 1L, 3L, 
2L, 4L, 3L, 2L, 3L, 1L, 5L, 5L, 7L, 1L, 3L, 3L, 2L, 1L, 3L, 4L, 
5L, 1L, 1L, 3L, 3L, 3L, 5L, 3L, 6L, 4L, 3L, 1L, 3L, 5L, 7L, 1L, 
3L, 4L, 5L, 3L, 3L, 1L, 1L, 1L, 7L, 3L, 1L, 4L, 3L, 3L, 5L, 1L, 
4L, 5L, 4L, 2L, 5L, 3L, 1L, 1L, 5L, 4L, 7L, 5L, 2L, 2L, 5L, 3L, 
1L, 1L, 2L, 3L, 5L, 3L, 7L, 5L, 1L, 5L, 3L, 1L, 1L, 1L, 1L, 7L, 
5L, 7L, 3L, 1L, 5L, 7L, 6L, 3L, 7L, 2L, 2L, 3L, 1L, 2L, 1L, 5L, 
5L, 2L, 4L, 1L, 1L, 2L, 1L, 4L, 7L, 3L, 2L, 5L, 3L, 2L, 4L, 2L, 
1L, 7L, 5L, 2L, 2L, 2L, 3L, 4L, 1L, 2L, 5L, 2L, 3L, 3L, 1L, 3L, 
2L, 3L, 5L, 1L, 3L, 1L, 5L, 4L, 5L, 4L, 5L, 5L, 5L, 1L, 3L, 3L, 
1L, 3L, 6L, 3L, 4L, 3L, 3L, 5L, 3L), Censor = c(0L, 1L, 1L, 0L, 
0L, 1L, 1L, 1L, 1L, 1L, 1L, 0L, 1L, 1L, 0L, 1L, 0L, 0L, 0L, 1L, 
1L, 0L, 0L, 0L, 1L, 1L, 0L, 0L, 1L, 0L, 0L, 1L, 1L, 1L, 0L, 0L, 
0L, 0L, 1L, 0L, 0L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 0L, 1L, 1L, 0L, 
1L, 0L, 0L, 1L, 1L, 1L, 1L, 1L, 0L, 1L, 1L, 1L, 1L, 1L, 1L, 0L, 
1L, 1L, 0L, 0L, 1L, 0L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 0L, 0L, 1L, 
1L, 1L, 1L, 1L, 1L, 0L, 1L, 0L, 1L, 0L, 0L, 1L, 0L, 0L, 1L, 1L, 
0L, 0L, 0L, 0L, 1L, 1L, 1L, 1L, 0L, 0L, 0L, 1L, 1L, 0L, 1L, 1L, 
0L, 0L, 1L, 0L, 1L, 1L, 0L, 1L, 0L, 0L, 0L, 0L, 0L, 1L, 0L, 1L, 
1L, 1L, 0L, 0L, 1L, 0L, 1L, 0L, 0L, 1L, 1L, 1L, 1L, 0L, 0L, 0L, 
0L, 1L, 1L, 0L, 0L, 0L, 1L, 1L, 0L, 1L, 0L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 0L, 0L, 1L, 1L, 0L, 1L, 0L, 0L, 0L, 1L, 1L, 1L, 0L, 0L, 
0L, 1L, 0L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 1L, 1L, 0L, 0L, 1L, 
0L, 1L, 1L, 1L, 0L, 1L, 0L, 1L, 1L, 0L, 1L, 1L, 0L, 1L, 1L, 0L, 
1L, 0L, 0L, 1L, 1L, 0L, 0L, 0L, 1L, 1L, 1L, 1L, 0L, 1L, 0L, 0L, 
1L, 0L, 1L, 1L, 0L, 1L, 1L, 1L, 0L, 1L, 0L, 1L, 0L, 0L, 0L, 1L, 
0L, 1L, 1L, 0L, 1L, 1L, 1L, 0L, 0L, 1L, 0L, 1L, 0L, 0L, 1L, 0L, 
1L, 1L, 0L, 1L, 0L, 1L, 0L, 1L, 1L, 0L, 1L, 1L, 1L, 0L, 1L, 0L, 
1L, 1L, 1L, 0L, 1L, 0L, 1L, 1L, 1L, 0L, 1L, 1L, 0L, 1L, 0L, 1L, 
1L, 0L, 0L, 0L, 0L, 0L, 1L, 0L)), class = "data.frame",
row.names = c(NA, -300L))

Here is the script:

library(mgcv)
library(ggplot2)

#Run the model
Model1 <- gam(SurvYear~
                (Gender)+
                s(Age, k=50)+  
                s(Highest_Educationmx, k=7),
              weights=Censor, data=df, gamma=1.5, family=cox.ph())
summary(Model1)

#Build a perspective chart
vis.gam(Model1, view=c("Age","Highest_Educationmx"),
        plot.type="persp", color="gray", se=-1, theta=45, phi=25,
        xlab="Age", ylab= "Highest Education",
        ticktype="detailed", zlim=c(-5.00, 2.00))

#Plot individual predictors using plot command from mgcv
plot(Model1, all.terms=T, rug=T, residuals=F, se=T, shade=T, seWithMean=T) 

#Plot individual predictors using ggplot instead of plot command from mgcv
#UNSURE HOW DO TO THIS

Answer 1

I'm biased (I wrote it) but you can use the gratia package for this.

You can use the draw() function as a replacement for plot.gam() , and if you want total control, just use evaluate_smooth() to produce a tidy representation of the smooth which is then easily plotted using ggplot2 .

Answer 2

Here is the script based on the suggestion from Gavin Simpson above:

library(gratia)

#Plot individual predictors using ggplot instead of the plot command from mgcv
sm <- gratia::evaluate_smooth(Model1, "Age") 
ggplot(sm, aes(x=Age, y=est)) + geom_line(size=1.0) +
  geom_ribbon(aes(ymax=est+se, ymin=est-se), alpha=0.20) +
  coord_cartesian(xlim=c(20.00, 75.00), ylim=c(-2.00, 1.00)) +
  scale_x_continuous(breaks=seq(20.00, 75.00, 5.00)) +
  scale_y_continuous(breaks=seq(-2.00, 1.00, 1.00)) +
  labs(title="Age") +
  xlab("Age") +
  ylab("Linear Risk Score") +
  theme(plot.title=element_text(size=10)) +
  geom_hline(yintercept=0, linetype="dashed", size=0.5) +
  geom_vline(xintercept=mean(df$Age), linetype="dashed", size=0.5)