How to avoid overlapp in geom_line

Question

I have a really long and extended code, which I have created with graphs that shows lines overlapping.

Could you please suggest something to disentangle this and make the reproduction clearer? Maybe some other options to modify the size of graph itself?

For example when save graphs with ggsave. Thanks

REPREX

Here the datset I am working on

df = structure(list(ANT = c("MEDIUM", "MEDIUM", "MEDIUM", "LOW", "LOW", 
"LOW", "HIGH", "HIGH", "HIGH"), BUT = c("flat", "tighted", "curved", 
"flat", "tighted", "curved", "flat", "tighted", "curved"), var = c("sighted", 
"sighted", "sighted", "sighted", "sighted", "sighted", "sighted", 
"sighted", "sighted"), mean = c(24.9795700007983, 22.5476640156841, 
24.4568444611505, 24.4608952261011, 22.179843452558, 25.1211893108135, 
24.1807279709165, 21.9700173259632, 25.5690587972475), low = c(24.1929993862564, 
22.0173613402898, 23.7793134256169, 23.5748441551253, 21.2709153846706, 
24.2111126105412, 23.376952383782, 21.3272260871708, 24.7053231240633
), up = c(25.7661406153403, 23.0779666910784, 25.1343754966842, 
25.3469462970768, 23.0887715204454, 26.0312660110858, 24.984503558051, 
22.6128085647556, 26.4327944704316)), row.names = c(NA, -9L), class = c("tbl_df", 
"tbl", "data.frame"))

the table where to get the statistics label from

val = structure(list(group1 = c("flat", "flat", "tighted", "flat", 
"flat", "tighted", "flat", "flat", "tighted"), group2 = c("tighted", 
"curved", "curved", "tighted", "curved", "curved", "tighted", 
"curved", "curved"), y.position = c(45.985, 48.4375, 50.89, 53.985, 
56.4375, 58.89, 58.493, 60.9455, 63.398), lab = c("0.0132  -  *", 
"0.585  -  ns", "0.0494  -  ns", "0.0826  -  ns", "0.618  -  ns", 
"0.0214  -  ns", "0.0458  -  ns", "0.211  -  ns", "0.00171  -  **"
)), row.names = c(NA, -9L), class = c("tbl_df", "tbl", "data.frame"
))

and the code that I have used

library(ggplot2) 
library(ggpubr)

ggplot(df, 
       aes(x= BUT, y= mean, 
           colour= ANT, group = ANT)) +
  
  geom_errorbar(aes(ymin= low, 
                    ymax= up), 
                colour="black", width=.10) +
  geom_line() + 
  geom_point(size= 3, shape= 21, fill="white") + 
  stat_pvalue_manual(
    val, 
    step.increase=0.05,label = "lab")

Two potential solutions (1) transform the y-values mathematically to spread them out or (2) adjust the y-axis limits to "zoom in" on the range. Easier to help if you [make this question reproducible](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) by including a small representative dataset in a plain text format and all the _relevant_ code. — neilfws, Apr 25 '23 at 23:52
Assuming you are using RStudio you can change the size of the displayed graph by resizing the "Plots" pane in RStudio, or by clicking "Zoom" and resizing the window. If you are saving the plot to disk then whatever save function you are using will let you specify dimensions. — Gregor Thomas, Apr 25 '23 at 23:52
ok, I have added the report as asked. So what I would like to do is make the lines as readable as possible. You could suggest even some commands to resize the graph because with RStudio even Zoom does not return the best reproduction either. What is important is that the white balls should reflect the position pointed out by the mean in the df dataset. And the should be for the error bars to, but in the complex should be better readable than this reprex — 12666727b9, Apr 26 '23 at 14:09

jared_mamrot · Accepted Answer · 2023-04-27T23:49:39.543

This is not how I would approach the problem - I would exclude non-significant interactions from the plot to 'tidy' things up - but here is an answer to your question:

library(tidyverse)
library(ggpubr)

df = structure(list(ANT = c("MEDIUM", "MEDIUM", "MEDIUM", "LOW", "LOW", 
                            "LOW", "HIGH", "HIGH", "HIGH"), BUT = c("flat", "tighted", "curved", 
                                                                    "flat", "tighted", "curved", "flat", "tighted", "curved"), var = c("sighted", 
                                                                                                                                       "sighted", "sighted", "sighted", "sighted", "sighted", "sighted", 
                                                                                                                                       "sighted", "sighted"), mean = c(24.9795700007983, 22.5476640156841, 
                                                                                                                                                                       24.4568444611505, 24.4608952261011, 22.179843452558, 25.1211893108135, 
                                                                                                                                                                       24.1807279709165, 21.9700173259632, 25.5690587972475), low = c(24.1929993862564, 
                                                                                                                                                                                                                                      22.0173613402898, 23.7793134256169, 23.5748441551253, 21.2709153846706, 
                                                                                                                                                                                                                                      24.2111126105412, 23.376952383782, 21.3272260871708, 24.7053231240633
                                                                                                                                                                       ), up = c(25.7661406153403, 23.0779666910784, 25.1343754966842, 
                                                                                                                                                                                 25.3469462970768, 23.0887715204454, 26.0312660110858, 24.984503558051, 
                                                                                                                                                                                 22.6128085647556, 26.4327944704316)), row.names = c(NA, -9L), class = c("tbl_df", 
                                                                                                                                                                                                                                                         "tbl", "data.frame"))
val = structure(list(group1 = c("flat", "flat", "tighted", "flat", 
                                "flat", "tighted", "flat", "flat", "tighted"), group2 = c("tighted", 
                                                                                          "curved", "curved", "tighted", "curved", "curved", "tighted", 
                                                                                          "curved", "curved"), y.position = c(45.985, 48.4375, 50.89, 53.985, 
                                                                                                                              56.4375, 58.89, 58.493, 60.9455, 63.398), lab = c("0.0132  -  *", 
                                                                                                                                                                                "0.585  -  ns", "0.0494  -  ns", "0.0826  -  ns", "0.618  -  ns", 
                                                                                                                                                                                "0.0214  -  ns", "0.0458  -  ns", "0.211  -  ns", "0.00171  -  **"
                                                                                                                              )), row.names = c(NA, -9L), class = c("tbl_df", "tbl", "data.frame"
                                                                                                                              ))
val <- val %>%
  mutate(lab = paste0(group1, " vs ", group2, ": ", lab))
  
ggplot(df, 
       aes(x= BUT, y= mean, 
           colour= ANT, group = ANT)) +
  
  geom_errorbar(aes(ymin= low, 
                    ymax= up), 
                colour="black", width=.10) +
  geom_line() + 
  geom_point(size= 3, shape= 21, fill="white") + 
  stat_pvalue_manual(val, step.increase=0.075,
                     label = "lab", size = 2.5,
                     y.position = 27)

^{Created on 2023-04-28 with reprex v2.0.2}

If you wanted to only plot the significant terms on the plot, you could do something like:

library(tidyverse)
library(ggpubr)

df = structure(list(ANT = c("MEDIUM", "MEDIUM", "MEDIUM", "LOW", "LOW", 
                            "LOW", "HIGH", "HIGH", "HIGH"), BUT = c("flat", "tighted", "curved", 
                                                                    "flat", "tighted", "curved", "flat", "tighted", "curved"), var = c("sighted", 
                                                                                                                                       "sighted", "sighted", "sighted", "sighted", "sighted", "sighted", 
                                                                                                                                       "sighted", "sighted"), mean = c(24.9795700007983, 22.5476640156841, 
                                                                                                                                                                       24.4568444611505, 24.4608952261011, 22.179843452558, 25.1211893108135, 
                                                                                                                                                                       24.1807279709165, 21.9700173259632, 25.5690587972475), low = c(24.1929993862564, 
                                                                                                                                                                                                                                      22.0173613402898, 23.7793134256169, 23.5748441551253, 21.2709153846706, 
                                                                                                                                                                                                                                      24.2111126105412, 23.376952383782, 21.3272260871708, 24.7053231240633
                                                                                                                                                                       ), up = c(25.7661406153403, 23.0779666910784, 25.1343754966842, 
                                                                                                                                                                                 25.3469462970768, 23.0887715204454, 26.0312660110858, 24.984503558051, 
                                                                                                                                                                                 22.6128085647556, 26.4327944704316)), row.names = c(NA, -9L), class = c("tbl_df", 
                                                                                                                                                                                                                                                         "tbl", "data.frame"))
val = structure(list(group1 = c("flat", "flat", "tighted", "flat", 
                                "flat", "tighted", "flat", "flat", "tighted"), group2 = c("tighted", 
                                                                                          "curved", "curved", "tighted", "curved", "curved", "tighted", 
                                                                                          "curved", "curved"), y.position = c(45.985, 48.4375, 50.89, 53.985, 
                                                                                                                              56.4375, 58.89, 58.493, 60.9455, 63.398), lab = c("0.0132  -  *", 
                                                                                                                                                                                "0.585  -  ns", "0.0494  -  ns", "0.0826  -  ns", "0.618  -  ns", 
                                                                                                                                                                                "0.0214  -  ns", "0.0458  -  ns", "0.211  -  ns", "0.00171  -  **"
                                                                                                                              )), row.names = c(NA, -9L), class = c("tbl_df", "tbl", "data.frame"
                                                                                                                              ))
val <- val %>%
  mutate(lab = paste0(group1, " vs ", group2, ": p=", lab)) %>%
  filter(str_detect(lab, "ns", negate = TRUE))
  
jitter_seed <- 123

ggplot(df, 
       aes(x = BUT, y = mean, 
           colour = ANT, group = ANT)) +
  geom_errorbar(aes(ymin = low, 
                    ymax = up),
                position = position_jitterdodge(jitter.width = 0.005, jitter.height = 0, seed = jitter_seed)) +
  geom_line(position = position_jitterdodge(jitter.width = 0.005, jitter.height = 0, seed = jitter_seed)) + 
  geom_point(size = 3, shape = 21, fill = "white",
             position = position_jitterdodge(jitter.width = 0.005, jitter.height = 0, seed = jitter_seed)) + 
  stat_pvalue_manual(val, step.increase = 0.075,
                     label = "lab", size = 3.5,
                     y.position = 27) +
  theme_bw(base_size = 16)

^{Created on 2023-04-28 with reprex v2.0.2}

Would this approach work for your use-case?

Wow, impressive. Thank you siìo much – 12666727b9 Apr 28 '23 at 14:30 — 12666727b9, Apr 28 '23 at 14:30

How to avoid overlapp in geom_line

1 Answers1