将带有线型的变量传递给 ggplot 线型 [英] Passing variable with line types to ggplot linetype

问题描述

我是 ggplot 的新手，所以请耐心等待.我正在绘制 35 个小区域地理区域的增长预测，即使使用了出色的 directlabels 库，这对于一个地块来说也是一个不健康的数量.但是我需要所有系列进行初步筛选.

I am new to ggplot so bear with me. I am charting out growth projections for 35 small-area geographies which is an unhealthy amount for one plot even with use of the fantastic directlabels library. However I need all the series for initial screening.

挑战在于使其具有可读性.我找到了@Ben Bolker 的修复方法，用于使用 大量不同的颜色，但我无法改变线型.35 系列不需要是唯一的，但我想使用 12 种不同的类型，使单个系列更易于阅读.

The challenge is to make it readable. I found a fix by @Ben Bolker for using large numbers of distinct colors but am having trouble varying the linetype. The 35 series don't need to be unique, but I would like to use the 12 different types to make individual series easier to read.

我的计划是创建一个随机列表，其中包含 12 种可能类型的 35 个元素，并将其作为 linetype 参数传递，但我无法使其正常工作，并出现错误:

My plan was to create a random list with 35 elements of the 12 possible types and pass that as the linetype argument, but I am having trouble getting it to work, with the error:

Error: Aesthetics must either be length one, or the same length as the dataProblems:lty

我的线型列表中有 35 个值.当然，我希望类型、颜色和所有内容都反映在图例中.

I have 35 values in the linetype list. Of course I would like for the types, colors and all to be reflected in the legend.

融化的数据是这样的；35 个系列中每个系列的 9 年观察:

The melted data looks like this; 9 years' observations for each of 35 series:

> simulation_long_index[16:24,]    
      year    geography    value
16    2018    sfr_2    101.1871
17    2019    sfr_2    101.1678
18    2020    sfr_2    101.2044
19    2012    sfr_3    100.0000
20    2013    sfr_3    100.1038
21    2014    sfr_3    100.2561
22    2015    sfr_3    100.0631
23    2016    sfr_3    100.8071
24    2017    sfr_3    101.2405    

这是我目前的代码:

lty <- data.frame(lty=letters[1:12][sample(1:12, 35,replace=T)])

g3<-ggplot(data=simulation_long_index,
   aes(
     x=as.factor(year), 
     y=value, 
     colour=geography,
     group=geography,
     linetype=lty$lty))+
       geom_line(size=.65) + 
       scale_colour_manual(values=manyColors(35)) +
     geom_point(size=2.5) +
     opts(title="growth")+
     xlab("Year") + 
     ylab(paste("Indexed Value (Rel. to 2012")) + 
     opts(axis.text.x=theme_text(angle=90, hjust=0))

print(g3)

添加

    scale_linetype_manual("",values=lty$lty) +

在 scale_color_manual 而不是 linetype 参数后生成图表，但线条都是相同的.那么，对于大系列计数，我如何使线条发生变化?

after scale_color_manual instead of the linetype argument produces the chart, but lines are all the same. How, then, do I get the lines to vary for large series counts?

推荐答案

使用 scale_..._manual 的技巧通常是发送一个命名向量作为 value争论.setNames 函数对此很有用

The trick with using scale_..._manual is often to send a named vector as the value argument. The setNames function is good for this

首先，一些虚拟数据

## some dummy data 
simulations<- expand.grid(year = 2012:2020, geography = paste0('a',1:35))
library(plyr)
library(RColorBrewer)
simulation_long_index <- ddply(simulations, .(geography), mutate, 
  value = (year-2012) * runif(1,-2, 2) + rnorm(9, mean = 0, sd = runif(1, 1, 3)))
## create a manyColors function 
manyColors <- colorRampPalette(brewer.pal(name = 'Set3',n=11))

接下来我们创建一个向量，它是从 1:12 开始的随机样本(带替换)，并将名称设置为与 geography 变量相同

Next we create a vector that is a random sample from 1:12 (with replacement) and set the names the same as the geography variable

lty <- setNames(sample(1:12,35,T), levels(simulation_long_index$geography))

这是它的样子

lty
## a1  a2  a3  a4  a5  a6  a7  a8  a9 a10 a11 a12 a13 a14 a15 a16 
## 7   5   8  11   2  10   3   2   5   4   6   6  11   8   2   2 
## a17 a18 a19 a20 a21 a22 a23 a24 a25 a26 a27 a28 a29 a30 a31 a32 
## 12   7   6   8  11   5   1   1   8  12   8   1  12   2   3   5 
## a33 a34 a35 
#7   1   3 

现在您可以将 line_type = geography 与 scale_linetype_manual(values = lty)

ggplot(data=simulation_long_index,
        aes(
          x=as.factor(year), 
          y=value, 
          colour=geography,
          group=geography,
          linetype = geography))+
            geom_line(size=.65) + 
            scale_colour_manual(values=manyColors(35)) +
            geom_point(size=2.5) +
            opts(title="growth")+
            xlab("Year") + 
            ylab(paste("Indexed Value (Rel. to 2012")) + 
            opts(axis.text.x=theme_text(angle=90, hjust=0)) +
            scale_linetype_manual(values = lty)

给你什么

顺便说一句，您真的想将年份绘制为因子变量吗?

As an aside, do you really want to plot the years as a factor variable?

这篇关于将带有线型的变量传递给 ggplot 线型的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！