ggplot2饼图：通过将ggrepel切片标签移向饼图的圆周来重新定位它们可能吗？
我正在尝试使用 ggplot2 创建以下 data.table obj 的饼图；
> pietable
cluster N P p
1: 1 962 17.4 8.70
2: 3 611 11.1 22.95
3: 10 343 6.2 31.60
4: 12 306 5.5 37.45
5: 8 290 5.2 42.80
6: 5 288 5.2 48.00
7: 7 259 4.7 52.95
8: 18 210 3.8 57.20
9: 4 207 3.7 60.95
10: 9 204 3.7 64.65
11: 16 199 3.6 68.30
12: 17 201 3.6 71.90
13: 14 174 3.1 75.25
14: 22 159 2.9 78.25
15: 6 121 2.2 80.80
16: 21 106 1.9 82.85
17: 2 101 1.8 84.70
18: 26 95 1.7 86.45
19: 11 89 1.6 88.10
20: 24 84 1.5 89.65
21: 32 71 1.3 91.05
22: 13 65 1.2 92.30
23: 38 50 0.9 93.35
24: 25 41 0.7 94.15
25: 36 36 0.7 94.85
26: 20 39 0.7 95.55
27: 28 33 0.6 96.20
28: 23 30 0.5 96.75
29: 31 24 0.4 97.20
30: 15 21 0.4 97.60
31: 34 22 0.4 98.00
32: 30 14 0.3 98.35
33: 33 19 0.3 98.65
34: 40 10 0.2 98.90
35: 29 11 0.2 99.10
36: 37 9 0.2 99.30
37: 19 8 0.1 99.45
38: 27 6 0.1 99.55
39: 39 6 0.1 99.65
40: 35 3 0.1 99.75
N 是特定簇的总数，P 是比例，p 是 P 的累积和。
用于创建饼图的 ggplot2 线如下；
ggplot(pietable, aes("", P)) +
geom_bar(
stat = "identity",
aes(
fill = rev(fct_inorder(cluster)))) +
geom_label_repel(
data = pietable[!P<1],
aes(
label = paste0(P, "%"),
y = p1,
#col = rev(fct_inorder(cluster))
),
point.padding = NA,
max.overlaps = Inf,
nudge_x = 1,
color="red",
force = 0.5,
force_pull = 0,
segment.alpha=0.5,
arrow=arrow(length = unit(0.05, "inches"),ends = "last", type = "open"),
show.legend = F
)+
geom_label_repel(
data = pietable[!P<1],
aes(label = cluster,
y = p1),
size=2,
col="black",
force = 0,
force_pull = 0,
label.size = 0.01,
show.legend = F
)+
scale_fill_manual(values = P40) +
coord_polar(theta = "y")+
theme_void()
这会生成一个像这样的饼图；
由于明显的原因，一些具有极小的值的饼图切片没有被标记。
我想做的是将切片标签（按数量）移向饼图的圆周来重新定位它们。
我想创建的饼图示例如下所示；
我很感激任何完成此任务的建议。
I am trying to create a pie chart of a following data.table obj by using ggplot2;
> pietable
cluster N P p
1: 1 962 17.4 8.70
2: 3 611 11.1 22.95
3: 10 343 6.2 31.60
4: 12 306 5.5 37.45
5: 8 290 5.2 42.80
6: 5 288 5.2 48.00
7: 7 259 4.7 52.95
8: 18 210 3.8 57.20
9: 4 207 3.7 60.95
10: 9 204 3.7 64.65
11: 16 199 3.6 68.30
12: 17 201 3.6 71.90
13: 14 174 3.1 75.25
14: 22 159 2.9 78.25
15: 6 121 2.2 80.80
16: 21 106 1.9 82.85
17: 2 101 1.8 84.70
18: 26 95 1.7 86.45
19: 11 89 1.6 88.10
20: 24 84 1.5 89.65
21: 32 71 1.3 91.05
22: 13 65 1.2 92.30
23: 38 50 0.9 93.35
24: 25 41 0.7 94.15
25: 36 36 0.7 94.85
26: 20 39 0.7 95.55
27: 28 33 0.6 96.20
28: 23 30 0.5 96.75
29: 31 24 0.4 97.20
30: 15 21 0.4 97.60
31: 34 22 0.4 98.00
32: 30 14 0.3 98.35
33: 33 19 0.3 98.65
34: 40 10 0.2 98.90
35: 29 11 0.2 99.10
36: 37 9 0.2 99.30
37: 19 8 0.1 99.45
38: 27 6 0.1 99.55
39: 39 6 0.1 99.65
40: 35 3 0.1 99.75
N is a total count of a particular cluster, P is a proportion , and p is a cumsum of P.
A ggplot2 line to create the pie chart is as follows;
ggplot(pietable, aes("", P)) +
geom_bar(
stat = "identity",
aes(
fill = rev(fct_inorder(cluster)))) +
geom_label_repel(
data = pietable[!P<1],
aes(
label = paste0(P, "%"),
y = p1,
#col = rev(fct_inorder(cluster))
),
point.padding = NA,
max.overlaps = Inf,
nudge_x = 1,
color="red",
force = 0.5,
force_pull = 0,
segment.alpha=0.5,
arrow=arrow(length = unit(0.05, "inches"),ends = "last", type = "open"),
show.legend = F
)+
geom_label_repel(
data = pietable[!P<1],
aes(label = cluster,
y = p1),
size=2,
col="black",
force = 0,
force_pull = 0,
label.size = 0.01,
show.legend = F
)+
scale_fill_manual(values = P40) +
coord_polar(theta = "y")+
theme_void()
This generates a pie chart like this;
Some of the pie slices with extremely small values are not labeled for an obvious reason.
What I'd like to do is to reposition slice labels (in number) by moving them toward circumference of the pie.
An example of the pie chart that I'd like to create looks like this;
I'd appreciate any suggestions to accomplish this.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
对于一个可视化来说，这是很多标签，您可能需要考虑不同的设计。
也就是说，您可以向相关几何图形添加
nudge_x = 0.33和segment.color = 'transparent'来调整标签的位置。请注意，为了使红色箭头与微移的黑色标签对齐，您必须将aes(after_stat(1.33))添加到红色几何图形，以匹配新的nudge_x黑色几何参数（+1.0）。这些行在下面的代码中标有# added注释。我对您的绘图代码做了一些轻微的修改（您的工作区中似乎有一些变量在您的帖子中不可用），但这是总体思路：
作为参考，数据我工作的集合：
This is a lot of labeling for one visualization, and you may want to consider a different design.
That said, you can add
nudge_x = 0.33andsegment.color = 'transparent'to the relevant geom to adjust the positioning of the labels. Note that in order to make the red arrows align to the nudged black labels, you have to addaes(after_stat(1.33))to the red geom, matching the newnudge_xparameter of the black geom (+1.0). These lines are marked with# addedcomments in the code below.I've made some light modifications to your plotting code (you seem to have variables in your workspace that are not available in your post), but this is the general idea:
And for reference, the data set I worked from: