linetype mapping in ggplot2 using geom_line()

Question

I have a dataset that looks like this:

   case prop weight  res
1     A   10    0.1 0.81
2     A   20    0.2 0.78
3     A   30    0.3 0.76
4     A   40    0.4 0.58
5     A   50    0.1 0.62
6     A   10    0.2 0.73
7     A   20    0.3 0.68
8     A   30    0.4 0.70
9     A   40    0.1 0.55
10    A   50    0.2 0.78
11    A   10    0.3 0.64
12    A   20    0.4 0.68
13    A   30    0.1 0.75
14    A   40    0.2 0.67
15    A   50    0.3 0.59
16    A   10    0.4 0.77
17    A   20    0.1 0.57
18    A   30    0.2 0.61
19    A   40    0.3 0.60
20    A   50    0.4 0.72
21    B   10    0.1 0.66
22    B   20    0.2 0.66
23    B   30    0.3 0.76
24    B   40    0.4 0.57
25    B   50    0.1 0.83
26    B   10    0.2 0.68
27    B   20    0.3 0.76
28    B   30    0.4 0.65
29    B   40    0.1 0.70
30    B   50    0.2 0.72
31    B   10    0.3 0.60
32    B   20    0.4 0.82
33    B   30    0.1 0.85
34    B   40    0.2 0.72
35    B   50    0.3 0.74
36    B   10    0.4 0.67
37    B   20    0.1 0.60
38    B   30    0.2 0.62
39    B   40    0.3 0.76
40    B   50    0.4 0.87
41    C   10    0.1 0.48
42    C   20    0.2 0.77
43    C   30    0.3 0.70
44    C   40    0.4 0.65
45    C   50    0.1 0.73
46    C   10    0.2 0.70
47    C   20    0.3 0.80
48    C   30    0.4 0.68
49    C   40    0.1 0.58
50    C   50    0.2 0.63
51    C   10    0.3 0.71
52    C   20    0.4 0.68
53    C   30    0.1 0.84
54    C   40    0.2 0.66
55    C   50    0.3 0.77
56    C   10    0.4 0.67
57    C   20    0.1 0.64
58    C   30    0.2 0.74
59    C   40    0.3 0.81
60    C   50    0.4 0.62

The data can be generated by the codes below:

 case = rep(c("A", "B", "C"), each=20)
    prop = rep(c("10", "20", "30", "40", "50"), 12)
    weight = as.factor(rep(c(0.1, 0.2, 0.3, 0.4), 15))
    res = round(rnorm(n=60, 0.7, 0.1), 2)

    dat = data.frame(case, prop, weight, res)
    dat

What I want to achieve is to have "prop" as x-axis, and "res" as y-axis, while using different colors to distinguish "case" and using different linetypes to distinguish "weight". For example, if weight=0.1, then use solid line; if weight=0.2, use dashed line, etc. From the codes below:

ggplot(data = dat, aes(x=prop, y=res, group=case, color=case)) +
  geom_line() +
  geom_point() + 
  theme_bw()

I can only get the following plot which is not desired...

I tried to add geom_line(aes(linetype=weight)), but an error shows

Error: geom_path: If you are using dotted or dashed lines, colour, size and linetype must be constant over the line

Is there a way to map "weight" to the linetype in ggplot2? Thank you!

erc · Accepted Answer · 2016-03-11 14:59:41Z

9

You could create a new variable to group by and to use for specifying the linetype..

dat$case.weight <- paste0(dat$case, dat$weight)

.. and add scale_linetype_manual():

ggplot(data = dat, aes(x=prop, y=res, group=case.weight, color=case, linetype=case.weight)) +
  scale_linetype_manual(values=rep(c("solid", "dashed", "dotted", "dotdash"),3), 
                        breaks=c("A0.1","A0.2", "A0.3", "A0.4"),
                        labels=c("0.1", "0.2", "0.3", "0.4"),
                        name="weight") +
  geom_line() +
  geom_point() + 
  theme_bw()

edited Mar 11, 2016 at 14:59

answered Mar 11, 2016 at 14:32

erc

10.2k11 gold badges62 silver badges93 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Jaap Over a year ago

Nice one. Maybe include facets as well. That will make the plot clearer imo.

alittleboy Over a year ago

@beetroot: Thank you very much! Just one quick question: is there a way to add a legend to show the different linetypes? I'd like readers to see that "solid", "dashed", "dotted", "dotdash" correspond to 0.1, 0.2, 0.3 and 0.4, respecitvely. Thanks!

erc Over a year ago

@alittleboy I edited my answer to include a legend for weight.

Collectives™ on Stack Overflow

linetype mapping in ggplot2 using geom_line()

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related