0

I have the following data frame:

     expected observed group
1: 0.5371429   0.0000     1
2: 1.3428571   1.3736     1
3: 2.6857143   2.4554     1
4: 5.3714286   3.6403     1
5: 0.5294118   0.0000     2
6: 1.3235294   1.1494     2
7: 2.6470588   1.1364     2
8: 5.2941176   4.9774     2
9: 0.5201207   0.0000     3
10: 1.3003018   1.4327    3
11: 2.6006036   2.5918    3
12: 5.2012072   8.0769    3
13: 0.5155039   1.4851    4
14: 1.2887597   1.0638    4
15: 2.5775194   3.1700    4
16: 5.1550388   6.2500    4
17: 0.4976959   0.0000    5
18: 1.2442396   1.2384    5
19: 2.4884793   3.1073    5
20: 4.9769585   4.8148    5

I would like to scatter plot each dataset according to group, so I have the following code:

sp <- ggplot(new_df, aes(x = expected, y = observed, colour = group)) + geom_point()

sp + scale_color_gradientn(colours = rainbow(5))

and receive the below plot: enter image description here

My question is how to add a linear line (intercept = 0,0) to each of the different groups? meaning, that in the end, I'll have 5 linear lines in different colors representing each group on the same plot.

And, is there a way to show the equation for each line in a legend?

Tung
  • 26,371
  • 7
  • 91
  • 115
Bella
  • 937
  • 1
  • 13
  • 25
  • have you look into `geom_abline`? – MLavoie Feb 21 '18 at 17:48
  • It will give one linear line, not for each group – Bella Feb 21 '18 at 18:15
  • provide a reproducible example. see also https://stackoverflow.com/questions/45624655/geom-abline-multiple-slopes-and-intercepts – MLavoie Feb 21 '18 at 18:15
  • You could try [`geom_smooth`](https://stackoverflow.com/q/12281335/3965651) for the lines and use a [model](https://stackoverflow.com/questions/31829528/specify-regression-line-intercept-r-ggplot2) to set the intercept – desc Feb 21 '18 at 18:18
  • It will also give one line for all. How to add the geom_smooth or geom_abline to get multiple lines according to each group? – Bella Feb 21 '18 at 18:22
  • Need a reproducible example. You say "a linear line to each of the different groups". This could mean that you want multiple linear lines to each point in a group creating a jagged line. Or it could mean that you want a single regression line for each group that doesn't actually go "to each of the different groups". Hint: We have to be able to see the desired output...you can sketch it, use paint, or be very clear in your description. You can also try aes(..., group = group, ...) – Adam Sampson Feb 21 '18 at 19:48

1 Answers1

2

You can get the linear lines and equation/R2 text with geom_smooth from ggplot2 and stat_poly_eq from ggpmisc package

    dat <- "expected    observed    group
    0.5371429   0   1
    1.3428571   1.3736  1
    2.6857143   2.4554  1
    5.3714286   3.6403  1
    0.5294118   0   2
    1.3235294   1.1494  2
    2.6470588   1.1364  2
    5.2941176   4.9774  2
    0.5201207   0   3
    1.3003018   1.4327  3
    2.6006036   2.5918  3
    5.2012072   8.0769  3
    0.5155039   1.4851  4
    1.2887597   1.0638  4
    2.5775194   3.17    4
    5.1550388   6.25    4
    0.4976959   0   5
    1.2442396   1.2384  5
    2.4884793   3.1073  5
    4.9769585   4.8148  5
    "  
    library(ggplot2)
    library(ggpmisc)

    df <- read.table(text = dat, header = TRUE)
    df$group <- factor(df$group)

    formula <- y ~ x # needed for ggpmisc's equation and R2 text

    # Put equation & R2 coef to the top left corner
    ggplot(df, aes(expected, observed, colour = group)) +
      geom_point(size = 2, alpha = 0.3) +
      geom_smooth(method = "lm", formula = formula, se = FALSE) +
      stat_poly_eq(aes(label = paste(..eq.label.., ..rr.label.., sep = "~~~~")), 
                   label.x.npc = "left", label.y.npc = "top",
                   formula = formula, parse = TRUE, size = 4) +
      scale_color_brewer(palette = "Dark2") +
      theme_bw(base_size = 16)

enter image description here

Tung
  • 26,371
  • 7
  • 91
  • 115