Plotting geom_point and geom_lines

Question

I am trying to make a multilines plot by using ggplot. I also want to plot over the lines some points (which mean the significance). The problem is that when I use geom_point, these points do not coincide with the line that I want. These would be the data:

dat1: 
                1-4      2-5     3-6      4-7     5-8     6-9    7-10        id
 mod1      -0.035930 0.121970 0.34689 0.034345 0.35312 0.52048 0.58536     mod1
 mod2      -0.094121 0.297150 0.37262 0.512140 0.63918 0.42127 0.73890     mod2
 mod3       0.810550 0.876070 0.57120 0.472640 0.67341 0.79332 0.80882     mod3
 mod4      -0.121970 0.010009 0.49783 0.920100 0.76192 0.45662 0.45526     mod4

dat2:
             1-4     2-5 3-6    4-7     5-8     6-9    7-10        id
 mod1         NaN     NaN NaN    NaN     NaN     NaN     NaN     mod1
 mod2         NaN     NaN NaN    NaN 0.63918     NaN     NaN     mod2
 mod3      0.81055 0.87607 NaN    NaN 0.67341 0.79332 0.80882    mod3
 mod5          NaN     NaN NaN 0.9201 0.76192     NaN     NaN    mod4

The plot will contain 4 lines and in the model with some values, I want to plot them with a point.

This is my try:

#Start Plotting
library(ggplot2)
library(reshape2)
dat_r$id <- nam_model  #names of models
dat_r1$id <- nam_model
df <- melt(dat_r,id='id')
df2 <-melt(dat_r1,id='id')

p <-ggplot(df, aes(x=variable,y=value, group=id)) +
     geom_line(aes(color=id), lwd=1) + geom_point(aes(x=df2$variable,y=df2$value,      group=df$id),size = 4)

Any suggestion?? I appreciate any idea!

Thanks in advance

Answer 1

Here's one possibility:

Read data, using check.names = FALSE because your variable names are not syntactically valid.

dat1 <- read.table(text = "                1-4      2-5     3-6      4-7     5-8     6-9    7-10        id
 mod1      -0.035930 0.121970 0.34689 0.034345 0.35312 0.52048 0.58536     mod1
 mod2      -0.094121 0.297150 0.37262 0.512140 0.63918 0.42127 0.73890     mod2
 mod3       0.810550 0.876070 0.57120 0.472640 0.67341 0.79332 0.80882     mod3
 mod4      -0.121970 0.010009 0.49783 0.920100 0.76192 0.45662 0.45526     mod4",
                   header = TRUE, check.names = FALSE)

dat2 <- read.table(text = "             1-4     2-5 3-6    4-7     5-8     6-9    7-10        id
 mod1         NaN     NaN NaN    NaN     NaN     NaN     NaN     mod1
 mod2         NaN     NaN NaN    NaN 0.63918     NaN     NaN     mod2
 mod3      0.81055 0.87607 NaN    NaN 0.67341 0.79332 0.80882    mod3
 mod5          NaN     NaN NaN 0.9201 0.76192     NaN     NaN    mod4",
                   header = TRUE, check.names = FALSE)

melt data to long format:

library(reshape2)
dat1m <- melt(dat1, id.var = "id")
dat2m <- melt(dat2, id.var = "id")

Plotting with data set for the lines, and another for the points:

library(ggplot2)
ggplot(data = df1m, aes(x = variable, y = value, colour = id, group = id)) +
  geom_line() +
  geom_point(data = df2m, size = 4)

A small note: in your aes call, avoid code like dataset$variable (e.g. df2$value). It may cause unwanted behaviour.

Answer 2

First read the data:

dat1 <- read.table(header=TRUE, check.names=FALSE, text="1-4      2-5     3-6      4-7     5-8     6-9    7-10        id
-0.035930 0.121970 0.34689 0.034345 0.35312 0.52048 0.58536     mod1
-0.094121 0.297150 0.37262 0.512140 0.63918 0.42127 0.73890     mod2
 0.810550 0.876070 0.57120 0.472640 0.67341 0.79332 0.80882     mod3
-0.121970 0.010009 0.49783 0.920100 0.76192 0.45662 0.45526     mod4")

dat2 <- read.table(header=TRUE, check.names=FALSE, text="1-4     2-5 3-6    4-7     5-8     6-9    7-10        id
 NaN     NaN NaN    NaN     NaN     NaN     NaN     mod1
 NaN     NaN NaN    NaN 0.63918     NaN     NaN     mod2
 0.81055 0.87607 NaN    NaN 0.67341 0.79332 0.80882    mod3
 NaN     NaN NaN 0.9201 0.76192     NaN     NaN    mod4")

Than transform the data into long format with the reshape2 package:

library(reshape2)
df1 <- melt(dat1, id="id")
df2 <- melt(dat2, id="id")

You can also use a combination of the dplyr and tidyr packages:

library(dplyr)
library(tidyr)

df1 <- dat1 %>% gather(var, value, 1:7)
df2 <- dat2 %>% gather(var, value, 1:7)

Binding the data together in one dataframe (which is not necessarily):

dat <- cbind(df1,df2[,3])
names(dat) <- c("id","var","value1","value2")

Finally create the plot:

ggplot(data=dat, aes(x=var, y=value1, color=id, group=id)) +
  geom_line(lwd=1) + 
  geom_point(aes(y=value2), size=4) +
  scale_x_discrete("\nModels") +
  scale_y_continuous("Value", breaks=c(0,0.2,0.4,0.6,0.8)) +
  theme_bw()

which gives:

When you don't want to bind the data together in one dataframe, you can use:

ggplot(data=df1, aes(x=var, y=value, color=id, group=id)) +
  geom_line(lwd=1) + 
  geom_point(data=df2, size=4) +
  scale_x_discrete("\nModels") +
  scale_y_continuous("Value", breaks=c(0,0.2,0.4,0.6,0.8)) +
  theme_bw()