Draw independent geom_smooth() lines for each level

[nuskcalb]

Hi there!

geom_smooth() seems to work either for all levels or for none. There is any way for geom_smooth() to draw the lines for the different levels independently of each other?

In the reprex below you can see that, for the first example, geom_smooth works for both "a" and "b" levels. In the second example, I use the formula = y ~ poly(x, 4), which requires 5 or more points to draw a line. This means "b" could show a line, and "a" couldn't, but nothing is shown.

Thanks in advance for the help!

library(tidyverse)

DF = data.frame(
  stringsAsFactors = FALSE,
              var1 = c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L),
              var2 = c(2L, 4L, 3L, 5L, 3L, 5L, 4L, 6L, 8L, 8L),
             var_g = c("a", "a", "a", "a", "b", "b", "b", "b", "b", "b")
)

# Works
DF %>% 
  ggplot(aes(var1, var2, color = var_g)) +
  geom_point() + 
  geom_smooth(method = "lm", formula = y ~ poly(x, 3), se = FALSE)

# Does not work
DF %>% 
  ggplot(aes(var1, var2, color = var_g)) +
  geom_point() + 
  geom_smooth(method = "lm", formula = y ~ poly(x, 4), se = FALSE)
#> Warning: Computation failed in `stat_smooth()`:
#> 'degree' must be less than number of unique points

^{Created on 2020-03-19 by the reprex package (v0.3.0)}

[dromano]

Not sure if this is what you meant:

library(tidyverse)

DF = data.frame(
stringsAsFactors = FALSE,
var1 = c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L),
var2 = c(2L, 4L, 3L, 5L, 3L, 5L, 4L, 6L, 8L, 8L),
var_g = c("a", "a", "a", "a", "b", "b", "b", "b", "b", "b")
)

# Works
DF %>% 
ggplot(aes(var1, var2, color = var_g)) +
geom_point() + 
geom_smooth(method = "lm", formula = y ~ poly(x, 3), se = FALSE)

# Works by passing data to 'geom_smooth()'
DF %>% 
ggplot(aes(var1, var2, color = var_g)) +
geom_point() + 
geom_smooth(
  data = DF %>% filter(var_g == 'b'),
  method = "lm", formula = y ~ poly(x, 4), se = FALSE)

^{Created on 2020-03-20 by the reprex package (v0.3.0)}

[nuskcalb]

That is a great idea! I slightly modified it so it will automatically filter depending on the number of points per group.

Thanks!

library(tidyverse)

DF = data.frame(
stringsAsFactors = FALSE,
var1 = c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L),
var2 = c(2L, 4L, 3L, 5L, 3L, 5L, 4L, 6L, 8L, 8L),
var_g = c("a", "a", "a", "a", "b", "b", "b", "b", "b", "b")
)

poly_n = 4
counts_filter = DF %>% count(var_g) %>% filter(n > poly_n)

DF %>%
ggplot(aes(var1, var2, color = var_g)) +
geom_point() +
geom_smooth(data = DF %>% filter(var_g %in% counts_filter),
method = "lm", formula = y ~ poly(x, poly_n), se = FALSE)