R图：在图外操作轴上的小数位数和图例大小

Question

Dearest Community! I have some problem with making my plot perfectly beautiful using R plot. I did a lot of research on this topic, but I was not able to solve the last remaining problems (furthermore I am very new to RStudio and coding in general). I have this plot plot_example which is created with this code:

df1 <- read.csv2("df1.csv", header=T, sep=";", dec=".")
el_x1 <- df1$element
el_y1 <- df1$ratio
df2 <- read.csv2("df2.csv", header=T, sep=";", dec=".")
el_x2 <- df2$element
el_y2 <- df2$ratio
df3 <- read.csv2("df3.csv", header=T, sep=";", dec=".")
el_x3 <- df3$element
el_y3 <- df3$ratio
df4 <- read.csv2("df4.csv", header=T, sep=";", dec=".")
el_x4 <- df4$element
el_y4 <- df4$ratio
df5 <- read.csv2("df5.csv", header=T, sep=";", dec=".")
el_x5 <- df5$element
el_y5 <- df5$ratio
df6 <- read.csv2("df6.csv", header=T, sep=";", dec=".")
el_x6 <- df6$element
el_y6 <- df6$ratio
df7 <- read.csv2("df7.csv", header=T, sep=";", dec=".")
el_x7 <- df7$element
el_y7 <- df7$ratio
df8 <- read.csv2("df8.csv", header=T, sep=";", dec=".")
el_x8 <- df8$element
el_y8 <- df8$ratio
df9 <- read.csv2("df9.csv", header=T, sep=";", dec=".")
el_x9 <- df9$element
el_y9 <- df9$ratio
df10 <- read.csv2("df10.csv", header=T, sep=";", dec=".")
el_x10 <- df10$element
el_y10 <- df10$ratio
df11 <- read.csv2("df11.csv", header=T, sep=";", dec=".")
el_x11 <- df11$element
el_y11 <- df11$ratio
par(fin = c (5,5), mar = c(5,5,6,1), xpd = "true")
plot (xlim = c(0.1,250), ylim = c(0.003,4), el_x1, el_y1, log = "xy", xlab = "element (ppm)", ylab = "ratio", pch = 0, cex = 0.7, col = "orange1", las = 1, cex.axis = 1, cex.lab = 1.8, yaxs = "i", xaxs = "i")
points (el_x2, el_y2, pch = 3, cex = 0.7, col = "olivedrab2", las = 1, yaxs = "i", xaxs = "i")
points (el_x3, el_y3, pch = 4, cex = 0.7, col = "darkseagreen1", las = 1, yaxs = "i", xaxs = "i")
points (el_x4, el_y4, pch = 2, cex = 0.7, col = "khaki1", las = 1, yaxs = "i", xaxs = "i")
points (el_x5, el_y5, pch = 1, cex = 0.7, col = "lavender", las = 1, yaxs = "i", xaxs = "i")
points (el_x6, el_y6, pch = 15, cex = 0.7, col = "seagreen", las = 1, yaxs = "i", xaxs = "i")
points (el_x7, el_y7, pch = 15, cex = 1, col = "indianred1", las = 1, yaxs = "i", xaxs = "i")
points (el_x8, el_y8, pch = 16, cex = 1, col = "royalblue1", las = 1, yaxs = "i", xaxs = "i")
points (el_x9, el_y9, pch = 15, cex = 1, col = "red2", las = 1, yaxs = "i", xaxs = "i")
points (el_x10, el_y10, pch = 16, cex = 1, col = "navy", las = 1, yaxs = "i", xaxs = "i")
points (el_x11, el_y11, pch = 16, cex = 1, col = "darkmagenta", las = 1, yaxs = "i", xaxs = "i")
legend ("topright", inset=c(-0.46,0), legend=c("df1verylongname", "df2verylongname", "df3verylongname", "df4verylongname", "df5verylongname", "df6verylongname", "df7verylongname","df8verylongname", "df9verylongname", "df10verylongname", "df11verylongname"), pch=c(0,3,4,2,1,15,15,15,16,16,16), col=c("orange1","olivedrab2","darkseagreen1","khaki1", "lavender", "seagreen", "indianred1", "red2", "royalblue2", "navy", "darkmagenta")) 

Now, I would like to make some modifikations.

1. (Most important) I have to manipulate the decimal digits on the axes (both). I want to keep 0.005, but I want to get rid of 50.0 (it should be: 50). I am not able yet to manipulate it with "sprintf" correctly. So how can I remove all .0xxx digits which are not necessary? Would be nice if there is another solution than manipulating the plot with an picture editing programm...
2. Can I adjust the size of the legend to the heigth of the plot? It should reach from the bottom to the top of the right box side (in this case from y value 0.003 to 4).
3. Is there another way to place the legend next to the plot than "xpd = true"? Because I would like to get rid of some data points which seems to be outliers. Of course I could adjust the size of the plot or delete the outliers, but it would be way easier if the points outside of the box are not shown in general/like before adding the xpd.
4. Is there a nice way to get more space between the axes digits and the axes label? It was hard work to get the plot to this size and my attempts to adjust the space destroyed everything, I don't really want to make the axes digits even more smaller... :( :(

Here is a little bit of the input data, I simplified the df to the necessary stuff:

sample  element ratio
1   0.6175  0.063568046
2   4.678133548 0.008924568
3   3.638120051 0.005707297
4   4.116887372 0.007867378
5   3.090387742 0.059814081
6   10.3098635  0.040600128
7   10.29649952 0.019962218
8   8.241753356 0.012910088
9   26.76850701 0.294731393
10  8.136495793 0.031161747
11  7.993317894 0.021593337
12  10.48361696 0.025802074
13  14.37169851 0.030482194
14  13.19394369 0.032786504
15  3.756483892 0.027182974
16  12.21848391 0.032594756
17  6.560291802 0.049019108
18  10.71912771 0.072430938
19  1.708155726 0.007771124
20  10.0684893  0.054408104
21  16.38043254 0.076730258
22  11.50283707 0.044073631
23  10.11554913 0.064254448
24  13.52364219 0.0707148
25  10.51594695 0.060657727
26  14.51045122 0.07431958
27  16.46248665 0.059498267
28  12.17582522 0.04927026
29  20.83513461 0.059772704
30  18.10278099 0.069570053
31  8.404223175 0.051083018
32  13.96869871 0.069607707
33  9.396932425 0.044899193
34  12.39775022 0.066800258
35  1.625963393 0.009627417
36  16.9934813  0.059799741
37  75.5    0.3775
38  73  0.350961538
39  45.3    0.191949153
40  92.5    0.409292035
41  39.8    0.176106195
42  48.2    0.194354839
43  48.9    0.188076923
44  52.6    0.219166667
45  44.4    0.1734375
46  53.9    0.245
47  168.4   0.559468439
48  165.8   0.587943262
49  87.9    0.412676056
50  99.5    0.328382838
51  105.9   0.318975904
52  172.7   0.579530201
53  190.4   0.602531646
54  206.6   0.586931818
55  101 0.321656051
56  101.8   0.329449838
57  117.7   0.346176471
58  97.9    0.337586207
59  90.5    0.285488959
60  182.1   0.550151057
61  109.9   0.399636364
62  8   0.053619303
63  4.9 0.039294306
64  4.7 0.035768645
65  6.3 0.041888298
66  8.3 0.035698925
67  6.7 0.043733681
68  25.2    0.193548387
69  24.7    0.168600683
70  59.1    0.226610429
71  30.6    0.234303216
72  45.1    0.223710317
73  93.2    0.315611243
74  83.1    0.309151786
75  88.8    0.311360449
76  34.8    0.169014085
77  36.6    0.183550652
78       
79  4.47    0.071565802
80  1.62    0.012413793
81  2.08    0.021010101
82  1.27    0.012330097
83  1.34    0.019705882
84  2.54    0.030274136
85  3.96    0.057641921
86  32.9    0.115438596
87  19.4    0.067595819
88  27.6    0.088461538
89  14.5    0.050699301
90  31  0.096875
91  36.3    0.078232759
92  27.8    0.072395833
93  8.56    0.039447005
94  10.1    0.040239044
95  4.44    0.056060606
96  3.94    0.05317139
97  4.18    0.059123055
98  5.23    0.07568741

Please contact if more input data is needed. I would be so happy if someone could help me in this case. Greetings!

Answer 1

Here is a solution with ggplot2 . It does not fulfill all of your requirements, because you've provided data for only 1 element, and I didn't fully understand all requirements. The axis labels are nonsensical, but you get the idea.

library(ggplot2)

df <- read.table(stringsAsFactors = FALSE, header = TRUE, text = "sample  element ratio
1   0.6175  0.063568046
2   4.678133548 0.008924568
3   3.638120051 0.005707297
4   4.116887372 0.007867378
5   3.090387742 0.059814081
6   10.3098635  0.040600128
7   10.29649952 0.019962218
8   8.241753356 0.012910088
9   26.76850701 0.294731393
10  8.136495793 0.031161747
11  7.993317894 0.021593337
12  10.48361696 0.025802074
13  14.37169851 0.030482194
14  13.19394369 0.032786504
15  3.756483892 0.027182974
16  12.21848391 0.032594756
17  6.560291802 0.049019108
18  10.71912771 0.072430938
19  1.708155726 0.007771124
20  10.0684893  0.054408104
21  16.38043254 0.076730258
22  11.50283707 0.044073631
23  10.11554913 0.064254448
24  13.52364219 0.0707148
25  10.51594695 0.060657727
26  14.51045122 0.07431958
27  16.46248665 0.059498267
28  12.17582522 0.04927026
29  20.83513461 0.059772704
30  18.10278099 0.069570053
31  8.404223175 0.051083018
32  13.96869871 0.069607707
33  9.396932425 0.044899193
34  12.39775022 0.066800258
35  1.625963393 0.009627417
36  16.9934813  0.059799741
37  75.5    0.3775
38  73  0.350961538
39  45.3    0.191949153
40  92.5    0.409292035
41  39.8    0.176106195
42  48.2    0.194354839
43  48.9    0.188076923
44  52.6    0.219166667
45  44.4    0.1734375
46  53.9    0.245
47  168.4   0.559468439
48  165.8   0.587943262
49  87.9    0.412676056
50  99.5    0.328382838
51  105.9   0.318975904
52  172.7   0.579530201
53  190.4   0.602531646
54  206.6   0.586931818
55  101 0.321656051
56  101.8   0.329449838
57  117.7   0.346176471
58  97.9    0.337586207
59  90.5    0.285488959
60  182.1   0.550151057
61  109.9   0.399636364
62  8   0.053619303
63  4.9 0.039294306
64  4.7 0.035768645
65  6.3 0.041888298
66  8.3 0.035698925
67  6.7 0.043733681
68  25.2    0.193548387
69  24.7    0.168600683
70  59.1    0.226610429
71  30.6    0.234303216
72  45.1    0.223710317
73  93.2    0.315611243
74  83.1    0.309151786
75  88.8    0.311360449
76  34.8    0.169014085
77  36.6    0.183550652
79  4.47    0.071565802
80  1.62    0.012413793
81  2.08    0.021010101
82  1.27    0.012330097
83  1.34    0.019705882
84  2.54    0.030274136
85  3.96    0.057641921
86  32.9    0.115438596
87  19.4    0.067595819
88  27.6    0.088461538
89  14.5    0.050699301
90  31  0.096875
91  36.3    0.078232759
92  27.8    0.072395833
93  8.56    0.039447005
94  10.1    0.040239044
95  4.44    0.056060606
96  3.94    0.05317139
97  4.18    0.059123055
98  5.23    0.07568741")

# 
ggplot(df, aes(element, ratio, fill="Element 1")) +
    geom_point(color="red2") +
    #geom_point(color=factor(1), size=1) 
    coord_cartesian( xlim=c(0, 300), ylim=c(0.0003, 4)) +
    scale_x_continuous(labels=c(0, "hundred", 200, ">=300"), breaks=c(0, 100, 200, 300)) +
    labs(x = "X axis", y = "Y axis (ratio/ppm)", fill = "Elements") +
    theme_minimal() +
    theme(#legend.position = "right", 
          legend.position = c(0.05, 0.95), 
          legend.justification = c(0.05, 0.95),
          axis.title.x = element_text(color="black", hjust = 0),
          axis.title.y = element_text(color="black", hjust = 0))

Answer 2

I can't really address items 2 through 4. However, the key to item 1 is creating the labels with signif and creating custom axes with axis .

First some sample data that spans several orders of magnitude:

set.seed(14)
x <- 10^(rnorm(30))
y <- 10^(rnorm(30))

Next choose where you want the axis ticks to lie and set up a vector to contain the labels.

tick_pos <- c(.03, .1, .3, 1, 3, 10, 30, 100, 300)
tick_lbl <- character(length(tick_pos))

Then calculate the appropriate label from the tick positions:

for (i in seq_along(tick_pos)) {
  tick_lbl[i] <- signif(tick_pos[i], 1)
}

Note that you can't just pass entire tick_pos vector into signif since the function forces the result to use a single format.

> signif(tick_pos, 1)
[1] 3e-02 1e-01 3e-01 1e+00 3e+00 1e+01 3e+01 1e+02 3e+02

Finally, create the plot without axes, add the custom axes, and then a frame:

plot(x, y, log = 'xy', axes = FALSE)
axis(1, labels = tick_lbl, at = tick_pos)
axis(2, labels = tick_lbl, at = tick_pos)
box()