Removing Zeros from GGplot

Question

I currently have the below plot in R. I was wondering if it was possible to end the curves once they hit zero. For example, once z hits zero is it possible to remove the line and points on the plot after 10^4?

ydot.dplyr <- ydot.n0 %>% select(variables, variable, value) %>%
                group_by(variables) %>%
                do(as.data.frame(spline(x= .[["variable"]], y= .[["value"]], n = nrow(.)*10)))


ggplot(ydot.n0, aes(variable, value)) +
   geom_point(aes(colour = factor(variables))) +
   geom_line(data = ydot.dplyr, aes(x = x, y = y, colour = factor(variables)))

   variables variable        value
1          1     10^2   0.00000000
2          x     10^2   0.00000000
3          y     10^2  -2.98409634
4          z     10^2  -7.92656901
5         xx     10^2   0.78848436
6         xy     10^2   0.00000000
7         xz     10^2   0.00000000
8         yy     10^2   0.00000000
9         yz     10^2   0.00000000
10         1   10^2.5   0.00000000
11         x   10^2.5  -2.28064813
12         y   10^2.5   4.00588146
13         z   10^2.5 -10.72326346
14        xx   10^2.5   0.61528135
15        xy   10^2.5           NA
16        xz   10^2.5           NA
17        yy   10^2.5           NA
18        yz   10^2.5           NA
19         1     10^3   0.00000000
20         x     10^3  -0.01460549
21         y     10^3   5.26084263
22         z     10^3  -2.95040569
23        xx     10^3   0.00000000
24        xy     10^3   0.00000000
25        xz     10^3   0.00000000
26        yy     10^3   0.63388822
27        yz     10^3  -0.59157518
28         1   10^3.5   0.00000000
29         x   10^3.5  28.00000000
30         y   10^3.5  -1.00000000
31         z   10^3.5   0.00000000
32        xx   10^3.5   0.00000000
33        xy   10^3.5   0.00000000
34        xz   10^3.5  -1.00000000
35        yy   10^3.5   0.00000000
36        yz   10^3.5   0.00000000
37         1     10^4   0.00000000
38         x     10^4  28.00000000
39         y     10^4  -1.00000000
40         z     10^4   0.00000000
41        xx     10^4   0.00000000
42        xy     10^4   0.00000000
43        xz     10^4  -1.00000000
44        yy     10^4   0.00000000
45        yz     10^4   0.00000000
46         1   10^4.5   0.00000000
47         x   10^4.5  28.00000000
48         y   10^4.5  -1.00000000
49         z   10^4.5   0.00000000
50        xx   10^4.5   0.00000000
51        xy   10^4.5   0.00000000
52        xz   10^4.5  -1.00000000
53        yy   10^4.5   0.00000000
54        yz   10^4.5   0.00000000
55         1     10^5   0.00000000
56         x     10^5  28.00000000
57         y     10^5  -1.00000000
58         z     10^5   0.00000000
59        xx     10^5   0.00000000
60        xy     10^5   0.00000000
61        xz     10^5  -1.00000000
62        yy     10^5   0.00000000
63        yz     10^5   0.00000000

I've edited and added data. By "hitting zero" I mean that the data points above at zero should no longer appear on the graph. Further, the data should only appear once it becomes nonzero.

welcome to SO. This question lacks a reproducible example (your sample data). Also, please clarify what you mean with 'once it hits zero'. How close to zero? This could be more challenging than you think - ever heard of floating numbers? I guess you want to remove all values after the first occurrence of zero? Something like `ydot.n0 %>% group_by(variable) %>% slice(1:which.min(value == 0))` may help — tjebo, Mar 19 '20 at 10:49
Sorry about that, I added some data from using: ydot.n0 <- melt(iterative.ydot.df, id.vars="variables") ydot.n0[ydot.n0$values == 0] <- NA The curved graph was later created using Dplyr — AW27, Mar 19 '20 at 11:26
Please take a look at [How to make a great R reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example). It is difficult to assist you if we cannot reproduce your results. — hendrikvanb, Mar 19 '20 at 11:59

tjebo · Answer 1 · 2020-03-19T12:49:29.597

Still not quite clear when you want to cut off the lines. Here is one suggestion.

library(tidyverse)

# fill NAs with 0, and convert x to numeric. This is for the spline calculation
ydot_filled <- ydot.n0 %>% mutate(value = zoo::na.fill(value, 0), x= as.numeric(as.factor(variable))) 

# I've changed your do call to lapply after splitting the data frame first, then binding back to create a new data frame. 
y_line <- lapply(split(ydot_filled, ydot.n0$variables), 
       function(u) as.data.frame(spline(x= u[["x"]], y= u[["value"]], n = nrow(u)*10))) %>%
  bind_rows(.id = 'variable') %>% 
  mutate(y = round(y,2)) %>% # helps avoid floating numbers
  group_by(variable) %>% 
  mutate(cum0 = cumsum(y == 0 )) %>% # this is to get the first 0
  filter(cum0 < 1) # remove all rows after first zero. Note some have value == 0 from the start

ggplot() +
  geom_point(data = ydot_filled, aes(x= x, y = value, colour = variables))  +
  geom_line(data = y_line, aes(x = x, y = y, colour = variable))

As you can see - it does not cut of some lines that 'cross' zero, because their splines never equal zero. this is what I meant with your problem is probably more challenging than you think. You'd need to also remove lines when the values change their symbol etc.

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

Data

# devtools::install_github("alistaire47/read.so")
ydot.n0 <- read.so::read.so("variables variable        value
1          1     10^2   0.00000000
2          x     10^2   0.00000000
3          y     10^2  -2.98409634
4          z     10^2  -7.92656901
5         xx     10^2   0.78848436
6         xy     10^2   0.00000000
7         xz     10^2   0.00000000
8         yy     10^2   0.00000000
9         yz     10^2   0.00000000
10         1   10^2.5   0.00000000
11         x   10^2.5  -2.28064813
12         y   10^2.5   4.00588146
13         z   10^2.5 -10.72326346
14        xx   10^2.5   0.61528135
15        xy   10^2.5           NA
16        xz   10^2.5           NA
17        yy   10^2.5           NA
18        yz   10^2.5           NA
19         1     10^3   0.00000000
20         x     10^3  -0.01460549
21         y     10^3   5.26084263
22         z     10^3  -2.95040569
23        xx     10^3   0.00000000
24        xy     10^3   0.00000000
25        xz     10^3   0.00000000
26        yy     10^3   0.63388822
27        yz     10^3  -0.59157518
28         1   10^3.5   0.00000000
29         x   10^3.5  28.00000000
30         y   10^3.5  -1.00000000
31         z   10^3.5   0.00000000
32        xx   10^3.5   0.00000000
33        xy   10^3.5   0.00000000
34        xz   10^3.5  -1.00000000
35        yy   10^3.5   0.00000000
36        yz   10^3.5   0.00000000
37         1     10^4   0.00000000
38         x     10^4  28.00000000
39         y     10^4  -1.00000000
40         z     10^4   0.00000000
41        xx     10^4   0.00000000
42        xy     10^4   0.00000000
43        xz     10^4  -1.00000000
44        yy     10^4   0.00000000
45        yz     10^4   0.00000000
46         1   10^4.5   0.00000000
47         x   10^4.5  28.00000000
48         y   10^4.5  -1.00000000
49         z   10^4.5   0.00000000
50        xx   10^4.5   0.00000000
51        xy   10^4.5   0.00000000
52        xz   10^4.5  -1.00000000
53        yy   10^4.5   0.00000000
54        yz   10^4.5   0.00000000
55         1     10^5   0.00000000
56         x     10^5  28.00000000
57         y     10^5  -1.00000000
58         z     10^5   0.00000000
59        xx     10^5   0.00000000
60        xy     10^5   0.00000000
61        xz     10^5  -1.00000000
62        yy     10^5   0.00000000
63        yz     10^5   0.00000000")

score 0 · Answer 2 · edited Jun 20 '20 at 09:12

I cannot quite use your dataset easily, but from what I see, you want to only plot those points that are > (or equal to) 0, right? In this case, it might have a pretty simple solution for you. I'll use a dummy dataset for the +/- y values:

df1 <- data.frame(x=seq(4, 10, by=0.1), y=cos(seq(4,10,by=0.1)), id='cos')
df2 <- data.frame(x=seq(4, 10, by=0.1), y=sin(seq(4,10,by=0.1)), id='sin')
df <- rbind(df1, df2)

Plotting these you get:

ggplot(df, aes(x,y)) + geom_line(aes(color=id))

If you want to eliminate any points below y=0, you just plot a subset of the data:

ggplot(df, aes(x,y)) + geom_line(data=subset(df,y>=0), aes(color=id))

Limits of x and y axes can be changed with xlim() and ylim() calls to your plot.

Removing Zeros from GGplot

2 Answers2