Scale ticks/ breaks to percentile in continuous colourbar in scale_fill_gradientn (ggplot2)

huangapple
go评论81阅读模式
英文:

Scale ticks/ breaks to percentile in continuous colourbar in scale_fill_gradientn (ggplot2)

问题

这是我使用ggplot制作的可重现的代码，生成了类似的图形：

  1. set.seed(123)
  2. library(ggplot2)
  3. df = data.frame(
  4. 	"x_var" = factor(rep(letters[1:10], 100)),
  5. 	"y_var" = factor(rep(LETTERS[1:10], each=100)),
  6. 	"fill_var" = sample(c(rexp(200, rate = 100),rexp(600, rate = 10), rexp(100, rate = 1), rexp(100, rate = 0.01)))
  7. )
  9. fill_colours = c("chartreuse4", "chartreuse3", "darkolivegreen2",
  10.                  "khaki", "gold", "darkorange", "firebrick", "darkred")
  11. fill_values_quantiles = seq(from = 0.05,
  12.                             to = 0.95,
  13.                             length.out = length(fill_colours) - 2)
  15. ggplot(data = df, aes(x = x_var, y = y_var, fill = fill_var)) +
  16. 	geom_tile() +
  17. 	scale_fill_gradientn(
  18. 		colours = fill_colours,
  19. 		values = c(0,
  20. 							 quantile(df$fill_var, fill_values_quantiles),
  21. 							 ceiling(max(df$fill_var)))
  22. 		)

这个代码生成了与您提供的图形相似的图。

您想要的图例条是使用scale_fill_gradientn函数的参数values来控制的。要获得与您提供的示例图例相似的效果，您可以按照以下方式修改values参数：

  1. values = c(0, quantile(df$fill_var, fill_values_quantiles), ceiling(max(df$fill_var))

这将在图例中创建与数据中的值对应的标记，这些值由values参数定义，它们将均匀分布在颜色条上。这将为您提供一个类似于您示例中的图例的连续颜色条。

如果您不想使用分位数，而是想使用某种转换，例如对数转换，您可以使用trans参数来指定转换函数。在这种情况下，您需要确保图例上的标记显示原始值而不是转换后的值。

英文:

Here's a reproducible code of the type of plot that I have with ggplot:

  1. set.seed(123)
  2. library(ggplot2)
  3. df = data.frame(
  4. 	"x_var" = factor(rep(letters[1:10], 100)),
  5. 	"y_var" = factor(rep(LETTERS[1:10], each=100)),
  6. 	"fill_var" = sample(c(rexp(200, rate = 100),rexp(600, rate = 10), rexp(100, rate = 1), rexp(100, rate = 0.01)))
  7. )
  9. fill_colours = c("chartreuse4", "chartreuse3", "darkolivegreen2",
  10.                  "khaki", "gold", "darkorange", "firebrick", "darkred")
  11. fill_values_quantiles = seq(from = 0.05,
  12.                             to = 0.95,
  13.                             length.out = length(fill_colours) - 2)
  15. ggplot(data = df, aes(x = x_var, y = y_var, fill = fill_var)) +
  16. 	geom_tile() +
  17. 	scale_fill_gradientn(
  18. 		colours = fill_colours,
  19. 		values = c(0,
  20. 							 quantile(df$fill_var, fill_values_quantiles),
  21. 							 ceiling(max(df$fill_var)))
  22. 		)

This gives this figure:

Scale ticks/ breaks to percentile in continuous colourbar in scale_fill_gradientn (ggplot2)

How do I get the legend bar like this?

Scale ticks/ breaks to percentile in continuous colourbar in scale_fill_gradientn (ggplot2)

I did this plot with this colour gradient because using quantiles to highlight the differences in the tile plots fits what I want to show. This plot will be done many times, with different dataframes.

I want the legend like the example I put above so that we can have a better idea of what's going on for the green values. With words, I want the ticks that correspond to the values I have put in the argument values of scale_fill_gradientn, and I want them to be evenly spaced. I still want a continuous colorbar.

Another idea instead of using quantiles: I could also probably work with a transformation of some sort, e.g. a log transformation, but in that case I'd want the real values of fill_val printed in the legend text at the ticks, and not the transformed values. But I'd like the other possibility better.

答案1

得分: 1

我想知道你是否想记录颜色比例尺？因为你有很少的大数值：

  1. ggplot(data = df, aes(x = x_var, y = y_var, fill = fill_var)) +
  2.   geom_tile() +
  3.   scale_fill_gradientn(colours = fill_colours,
  4.                         trans="log10",
  5.                         breaks=trans_breaks('log10', function(x) 10^x),
  6.                         labels=trans_format('log10', function(x) round(10^x, 2)))

[![enter image description here][1]][1]

  2. <details>
  3. <summary>英文:</summary>
  5. I wonder if you want to log teh color scale? since you have very few large values:
  7.     ggplot(data = df, aes(x = x_var, y = y_var, fill = fill_var)) +
  8.       geom_tile() +
  9.    
  10.       scale_fill_gradientn(colours = fill_colours,
  11.                             trans=&quot;log10&quot;,
  12.                             breaks=trans_breaks(&#39;log10&#39;, function(x) 10^x),
  13.                             labels=trans_format(&#39;log10&#39;, function(x) round(10^x,2)))
  15. [![enter image description here][1]][1]
  18.   [1]: https://i.stack.imgur.com/sqQit.png
  20. </details>
  24. # 答案2
  25. **得分**: 1
  27. 你已经非常接近了。你的问题基本上是“如何将我的填充向量转换为百分位数”。为此，你可以使用 `ecdf` 函数。代码中的其他注释。
  29. ```r
  30. ## 稍作更改以包括0和1
  31. fill_values_quantiles <- seq(0, 1, length.out = length(fill_colours))
  32. ## 使用这个向量来表示标签的分位数断点（！）
  33. quants <- quantile(df$fill_var, fill_values_quantiles)
  34. ## 将填充中的每个值转换为分位数
  35. df$ptile_var <- ecdf(df$fill_var)(df$fill_var)
  37. ## 使用分位数来填充
  38. ggplot(data = df, aes(x = x_var, y = y_var, fill = ptile_var)) +
  39.   geom_tile() +
  40.   scale_fill_gradientn(
  41.     ## 使用上面的向量来设置断点和标签
  42.     colours = fill_colours,
  43.     breaks = fill_values_quantiles,
  44.     labels = round(quants, 3)
  45.   )

Scale ticks/ breaks to percentile in continuous colourbar in scale_fill_gradientn (ggplot2)

英文:

You're very close. Your question is basically "how to convert my fill vector into percentiles". For this you can use the ecdf function. Other comments in the code.

  1. ## slightly changed to include 0 and 1
  2. fill_values_quantiles &lt;- seq(0, 1, length.out = length(fill_colours))
  3. ## use this for a vector of your quantile breaks for the labels (!)
  4. quants &lt;- quantile(df$fill_var, fill_values_quantiles)
  5. ## convert every value in your fill to quantiles
  6. df$ptile_var &lt;- ecdf(df$fill_var)(df$fill_var)
  8. ## use the percentiles for fill
  9. ggplot(data = df, aes(x = x_var, y = y_var, fill = ptile_var)) +
  10.   geom_tile() +
  11.   scale_fill_gradientn(
  12. ## use your vectors from above for breaks and labels 
  13.     colours = fill_colours,
  14.     breaks = fill_values_quantiles,
  15.     labels = round(quants, 3)
  16.   )

Scale ticks/ breaks to percentile in continuous colourbar in scale_fill_gradientn (ggplot2)<!-- -->

huangapple
  • 本文由 发表于 2023年3月31日 22:56:02
  • 转载请务必保留本文链接：https://go.coder-hub.com/75899955.html
  • geom-tile
  • ggplot2
  • gradient
  • legend
  • r
匿名

发表评论

匿名网友

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen:

确定