Python 中的 Violin Plot 在对数尺度上的问题 [英] Violin Plot troubles in Python on log scale

问题描述

在我的绘图上使用对数刻度时，我的小提琴绘图显示出奇怪的格式.我已经尝试过使用matplotlib和seaborn来获得类似的结果.

My violin plots are showing weird formats when using a log scale on my plots. I've tried using matplotlib and seaborn and I get very similar results.

import matplotlib.pyplot as plt
import seaborn as sns

data = [[1e-05, 0.00102, 0.00498, 0.09154, 0.02009, 1e-05, 0.06649, 0.42253, 0.02062, 0.10812, 0.07128, 0.03903, 0.00506, 0.13391, 0.08668, 0.04127, 0.00927, 0.00118, 0.063, 0.18392, 0.05948, 0.07774, 0.14018, 0.0133, 0.00339, 0.00271, 0.05233, 0.00054, 0.0593, 1e-05, 0.00076, 0.03409, 0.71491, 0.02311, 0.10246, 0.12491, 0.05164, 0.1553, 0.01079, 0.01734, 0.02239, 0.1347, 0.02877, 0.04752, 0.00333, 0.04553, 0.03189, 0.00947, 0.00158, 0.00888, 0.12663, 0.07531, 0.12367, 0.11346, 0.06638, 0.06154, 1e-05, 0.1838, 0.08659, 0.05654, 0.07658, 0.0348, 0.02954, 0.0123, 0.01529, 0.05559, 0.00416, 0.00038, 0.14142, 0.00164, 0.03671, 0.10609, 0.01209, 0.0024, 0.11718, 0.11224, 0.06032, 0.09632, 0.12216, 0.00087, 0.06746, 0.00433, 0.06836, 0.09928, 2e-05, 0.14116, 0.05718, 0.01196, 0.04297, 0.00709, 0.10535, 0.04772, 0.05691, 0.06277, 1e-05, 0.03917, 0.0026, 0.06763, 0.02083, 0.32244, 0.00561, 0.03399, 0.08146, 0.10606, 0.01482, 0.00339, 0.02275, 0.00685, 0.1536, 0.0592, 0.08869, 1e-05, 0.20489, 0.00094, 0.00714, 0.06355, 0.03414, 0.03002, 0.02365, 0.04376, 0.0246, 0.02745, 0.07604, 0.12069, 1e-05, 0.02974, 0.10681, 0.00987, 0.02543, 0.01416, 0.00098, 3e-05, 0.00967, 0.11958, 0.02882, 0.03634, 0.19232, 0.12058, 0.36535, 0.07428, 0.02829, 0.09189, 0.03677, 0.00036, 0.0463, 0.57029, 0.0105, 0.00015, 0.06212, 0.0329, 0.06102, 0.12267], 
[0.01219, 0.14638, 0.03822, 0.05784, 0.03615, 0.03288, 0.00986, 0.05331, 0.01434, 0.00999, 0.05272, 0.03269, 0.0682, 0.15455, 0.09675, 0.02272, 0.0027, 0.01955, 0.06194, 0.00115, 0.07799, 0.03987, 0.11152, 0.07229, 0.007, 0.00075, 0.04499, 0.01534, 0.04301, 0.01247, 0.09511, 0.02297, 0.05538, 0.04614, 0.07359, 0.06909, 1e-05, 0.04247, 0.05485, 0.00071, 0.082, 0.07614, 0.03751, 0.01625, 0.03309, 0.03228, 0.08109, 0.02171, 0.07246, 0.00353, 0.02434, 0.01394, 0.037, 0.02429, 0.15162, 0.0527, 0.0201, 0.07954, 0.07626, 0.09285, 0.05071, 0.01224, 0.06331, 0.07556, 0.04952, 0.00052, 0.00588, 0.132, 0.00067, 0.00012, 0.00084, 0.03865, 0.02362, 0.08976, 0.18545, 0.04882, 0.03789, 0.05006, 0.02979, 0.003, 0.09262, 0.05668, 0.02486, 0.05855, 0.11588, 0.07713, 0.10428, 0.00706, 0.02467, 0.13257, 0.11547, 0.06143, 0.09478, 0.06099, 0.02483, 0.09312, 0.16867, 0.07236, 0.10962, 0.04149, 0.05005, 0.09087, 0.0313, 0.03697, 0.07201, 2e-05, 0.00259, 0.00115, 0.03907, 0.02931, 0.14907, 0.05598, 0.07087, 0.09709, 0.10653, 0.11936, 0.08196, 0.1213, 0.00627, 0.08496, 0.00038, 0.03537, 0.20043, 0.05159, 0.05872, 0.07754, 0.07621, 0.05924, 0.09587, 0.02653, 0.07135, 1e-05, 0.01377, 0.0062, 0.01965, 0.00115, 0.07529, 0.04709, 0.05458, 0.10895, 0.02195, 0.04534, 0.015, 0.00577, 0.05784, 0.01691, 0.08103, 0.04178, 0.04328, 0.01204, 0.03463, 0.03805, 0.01231, 0.03646, 0.01162, 0.16536, 0.03471, 0.00541, 0.09088, 0.06447, 0.07263, 0.05924, 0.0952, 0.09938, 0.04464, 0.05543, 0.03827, 0.11514, 0.02803, 0.09589, 0.0254, 0.05351, 0.00171, 0.00856, 0.05828, 0.11975, 7e-05, 0.07093, 0.06077, 0.0384, 0.00163, 0.05992, 0.00463, 0.00975, 0.00429, 0.12965, 0.03388, 0.02372, 0.07622, 0.04341, 0.06637, 0.00578, 0.06946, 0.00469, 0.11668, 0.07033, 0.06806, 0.05505, 0.02195, 0.05089, 0.03404, 0.00552, 0.05331, 0.03695, 0.41581, 0.01553, 0.02045, 0.09779, 0.03842, 0.01115, 0.05392, 0.01147, 0.05855, 0.05588, 0.20745, 0.01536, 0.03993, 0.07677, 0.01388, 0.0029, 0.00235, 0.05823, 0.05237, 0.00425, 0.09225, 0.00703, 0.24038, 0.06733, 0.00064, 0.08959, 0.04365, 0.02308, 0.04566, 0.08395, 0.0038, 0.05322, 0.0145, 0.02012, 0.07084, 0.08202, 0.01091, 0.03738, 0.03798, 0.03473, 0.08534, 0.00133, 0.04046, 0.10119, 0.0317, 0.00312, 0.03614, 0.10442, 0.13286, 0.0042, 0.04229, 0.01735, 0.09879, 0.07516, 0.00303, 0.08062, 0.09347, 0.03473, 0.05099, 0.16373, 0.08988, 0.04696, 0.07488, 0.12159, 0.11098, 0.00549, 0.00122, 0.05276, 0.09883, 0.01346, 0.02059, 0.07394, 0.0413, 0.08766, 0.0124, 0.09913, 0.00754, 0.15671, 0.02699, 0.09978, 1e-05, 0.00243, 0.02819, 0.00027, 0.05793, 0.03165, 0.10168, 0.00042, 0.00044, 0.01332, 0.00542, 0.05946, 0.009, 0.10857, 0.01699, 1e-05, 0.00073, 0.10842, 0.17143, 0.00036, 0.00014, 0.10508, 0.01333, 0.34202, 0.12201, 0.04618, 0.02507, 0.02939, 0.03497, 0.01905, 0.00136, 0.02354, 0.00061, 0.08514, 0.14529, 0.04097, 0.12821, 0.18862], 
[0.04683, 0.02943, 0.07885, 0.07846, 0.06855, 0.02815, 0.00792, 0.0826, 0.00554, 0.01041, 0.03957, 0.0126, 0.08399, 0.15046, 0.15594, 0.03941, 0.0428, 0.11343, 0.15665, 0.07381, 0.04386, 0.12008, 0.04816, 0.04844, 0.08248, 0.08023, 0.03011, 0.00464, 0.07204, 0.08376, 0.05777, 0.06164, 0.00697, 0.02023, 0.04844, 0.0592, 0.00954, 0.06357, 0.0122, 0.05905, 0.00705, 0.0054, 0.08822, 0.06056, 0.02598, 0.02136, 0.05638, 0.03768, 0.05101, 0.08908, 0.0384, 0.01579, 0.04023, 0.03746, 0.17236, 0.08293, 0.12469, 0.14018, 0.04301, 0.07258, 0.02678, 0.08078, 0.07698, 0.06346, 0.06984, 0.04832, 0.07512, 0.0342, 0.05339, 0.026, 0.11585, 0.02744, 0.00979, 0.01312, 0.05915, 0.01326, 0.00107, 0.00737, 0.05971, 0.0451, 0.05788, 0.0007, 0.0043, 0.00142, 0.0019, 0.00055, 0.00223, 0.02441, 0.04555, 0.03869, 0.05791, 0.05517, 0.15743, 0.04517, 0.47114, 0.05639, 0.00152, 0.00371, 1e-05, 1e-05, 0.04192, 0.02758, 0.01945, 0.02763, 0.04021, 0.02844, 0.01823, 0.10665, 0.02067, 0.05433, 0.05591, 0.00733, 0.00858, 0.01949, 0.06519, 0.07793, 0.00199, 0.09916, 0.08717, 0.06273, 0.09408, 0.00638, 0.00248, 0.08922, 0.09157, 0.03525, 0.01791, 0.06016, 0.01939, 0.12194, 0.08303, 0.0831, 0.02714, 0.06312, 0.11584, 0.11334, 0.04314, 0.02575, 0.00629, 0.02408, 0.02274, 0.03037, 0.06737, 0.0175, 0.00888, 0.06568, 0.0839, 0.0085, 0.00831, 0.00154, 0.01072, 0.01289, 0.09074, 0.02131, 0.02997, 0.02343, 0.02355, 0.05324, 0.09564, 0.17995, 0.00828, 0.0148, 0.01858, 0.02106, 0.00288, 0.00344, 0.001, 0.02143, 0.00732, 0.01458, 0.01547, 0.01742, 0.00032, 0.24005, 0.00028, 0.00302, 0.07275, 0.04579, 0.06316, 0.02572, 0.09316, 0.03062, 0.10521, 0.07123, 0.03069, 0.07958, 0.04484, 0.01948, 0.01951, 0.01282, 0.00868, 0.07931, 0.01105, 0.01235, 0.09297, 0.06959, 0.00716, 0.0271, 0.00592, 0.09362, 0.00319, 0.00859, 0.08486, 0.02001, 0.00194, 0.04189, 0.09024, 0.07705, 0.07365, 0.01123, 0.03202, 0.01361, 0.00098, 0.00397, 0.00139, 0.00397, 0.00445, 1e-05, 0.00267, 0.06564, 0.06567, 0.06566, 0.06566, 0.09249, 0.03475, 0.0338, 0.0664, 0.02986, 0.04024, 0.00835, 0.04304, 0.04081, 0.04534, 0.06636, 0.03312, 0.06175, 0.03117, 0.02243, 0.03454, 0.11135, 0.07016, 0.0681, 0.09716, 0.02589, 0.4367, 0.08293, 0.11834, 0.00191, 0.10913, 0.00159, 0.0638, 0.01808, 0.00116, 0.00911, 0.01408, 0.09179, 0.02122, 0.05026, 0.05144, 0.03169, 0.06674]]

fig, ax = plt.subplots(1,3, sharey=True)
sns.violinplot(data=data, ax=ax[0])
sns.swarmplot(data=data, ax=ax[1])
sns.stripplot(data=data, ax=ax[2])

在线性范围内使用数据时，一切看起来都很好.但是，我的很多数据都在0.1到0.00001之间，因此我想使用对数刻度来更好地可视化.

When using the data on a linear scale, everything looks fine. However, a lot of my data is between 0.1 and 0.00001 so I wanted to use a log scale for better visualization.

切换到对数刻度时:

plt.yscale('log')
plt.ylim(0.000001, 1)

我的Swarmplot和Stripplot图看起来不错，但是，小提琴图并没有向底部凝结.请注意，我也没有任何负值，但是小提琴图总是表明我有负值.

My swarmplot and stripplot plots look fine, however, the violin plots do not condense towards the bottom. Notice that I also don't have any negative values, but the violin plots always suggest that I do.

总的来说，我本来希望我的小提琴图看起来更像这样(这是在 R 中完成的).

Overall, I would have expected my violin plots to look something more like this (which was done in R).

有关如何使用 seaborn 或 matplotlib 或其他基于 Python 的可视化使小提琴图更像最后一张图中的图(即在数据点较少时进行压缩)的任何建议?

Any suggestions on how to get the violin plots to act more like the plots in the last picture (i.e. condensing when there are fewer data points) using seaborn or matplotlib, or another python based visualization?

推荐答案

另一个可以尝试的选择是转换数据.

Another option to try, is to transform the data.

y 轴的刻度标签可以使用自定义格式器重写.并且可以生成类似于对数图的小刻度.

The tick labels for the y-axis can be rewritten using a custom formatter. And minor ticks similar to a log plot can be generated.

import matplotlib.pyplot as plt
from matplotlib import ticker as mticker
import seaborn as sns
import numpy as np

data = [[1e-05, 0.00102, 0.00498, 0.09154, 0.02009, 1e-05, 0.06649, 0.42253, 0.02062, 0.10812, 0.07128, 0.03903, 0.00506, 0.13391, 0.08668, 0.04127, 0.00927, 0.00118, 0.063, 0.18392, 0.05948, 0.07774, 0.14018, 0.0133, 0.00339, 0.00271, 0.05233, 0.00054, 0.0593, 1e-05, 0.00076, 0.03409, 0.71491, 0.02311, 0.10246, 0.12491, 0.05164, 0.1553, 0.01079, 0.01734, 0.02239, 0.1347, 0.02877, 0.04752, 0.00333, 0.04553, 0.03189, 0.00947, 0.00158, 0.00888, 0.12663, 0.07531, 0.12367, 0.11346, 0.06638, 0.06154, 1e-05, 0.1838, 0.08659, 0.05654, 0.07658, 0.0348, 0.02954, 0.0123, 0.01529, 0.05559, 0.00416, 0.00038, 0.14142, 0.00164, 0.03671, 0.10609, 0.01209, 0.0024, 0.11718, 0.11224, 0.06032, 0.09632, 0.12216, 0.00087, 0.06746, 0.00433, 0.06836, 0.09928, 2e-05, 0.14116, 0.05718, 0.01196, 0.04297, 0.00709, 0.10535, 0.04772, 0.05691, 0.06277, 1e-05, 0.03917, 0.0026, 0.06763, 0.02083, 0.32244, 0.00561, 0.03399, 0.08146, 0.10606, 0.01482, 0.00339, 0.02275, 0.00685, 0.1536, 0.0592, 0.08869, 1e-05, 0.20489, 0.00094, 0.00714, 0.06355, 0.03414, 0.03002, 0.02365, 0.04376, 0.0246, 0.02745, 0.07604, 0.12069, 1e-05, 0.02974, 0.10681, 0.00987, 0.02543, 0.01416, 0.00098, 3e-05, 0.00967, 0.11958, 0.02882, 0.03634, 0.19232, 0.12058, 0.36535, 0.07428, 0.02829, 0.09189, 0.03677, 0.00036, 0.0463, 0.57029, 0.0105, 0.00015, 0.06212, 0.0329, 0.06102, 0.12267],
[0.01219, 0.14638, 0.03822, 0.05784, 0.03615, 0.03288, 0.00986, 0.05331, 0.01434, 0.00999, 0.05272, 0.03269, 0.0682, 0.15455, 0.09675, 0.02272, 0.0027, 0.01955, 0.06194, 0.00115, 0.07799, 0.03987, 0.11152, 0.07229, 0.007, 0.00075, 0.04499, 0.01534, 0.04301, 0.01247, 0.09511, 0.02297, 0.05538, 0.04614, 0.07359, 0.06909, 1e-05, 0.04247, 0.05485, 0.00071, 0.082, 0.07614, 0.03751, 0.01625, 0.03309, 0.03228, 0.08109, 0.02171, 0.07246, 0.00353, 0.02434, 0.01394, 0.037, 0.02429, 0.15162, 0.0527, 0.0201, 0.07954, 0.07626, 0.09285, 0.05071, 0.01224, 0.06331, 0.07556, 0.04952, 0.00052, 0.00588, 0.132, 0.00067, 0.00012, 0.00084, 0.03865, 0.02362, 0.08976, 0.18545, 0.04882, 0.03789, 0.05006, 0.02979, 0.003, 0.09262, 0.05668, 0.02486, 0.05855, 0.11588, 0.07713, 0.10428, 0.00706, 0.02467, 0.13257, 0.11547, 0.06143, 0.09478, 0.06099, 0.02483, 0.09312, 0.16867, 0.07236, 0.10962, 0.04149, 0.05005, 0.09087, 0.0313, 0.03697, 0.07201, 2e-05, 0.00259, 0.00115, 0.03907, 0.02931, 0.14907, 0.05598, 0.07087, 0.09709, 0.10653, 0.11936, 0.08196, 0.1213, 0.00627, 0.08496, 0.00038, 0.03537, 0.20043, 0.05159, 0.05872, 0.07754, 0.07621, 0.05924, 0.09587, 0.02653, 0.07135, 1e-05, 0.01377, 0.0062, 0.01965, 0.00115, 0.07529, 0.04709, 0.05458, 0.10895, 0.02195, 0.04534, 0.015, 0.00577, 0.05784, 0.01691, 0.08103, 0.04178, 0.04328, 0.01204, 0.03463, 0.03805, 0.01231, 0.03646, 0.01162, 0.16536, 0.03471, 0.00541, 0.09088, 0.06447, 0.07263, 0.05924, 0.0952, 0.09938, 0.04464, 0.05543, 0.03827, 0.11514, 0.02803, 0.09589, 0.0254, 0.05351, 0.00171, 0.00856, 0.05828, 0.11975, 7e-05, 0.07093, 0.06077, 0.0384, 0.00163, 0.05992, 0.00463, 0.00975, 0.00429, 0.12965, 0.03388, 0.02372, 0.07622, 0.04341, 0.06637, 0.00578, 0.06946, 0.00469, 0.11668, 0.07033, 0.06806, 0.05505, 0.02195, 0.05089, 0.03404, 0.00552, 0.05331, 0.03695, 0.41581, 0.01553, 0.02045, 0.09779, 0.03842, 0.01115, 0.05392, 0.01147, 0.05855, 0.05588, 0.20745, 0.01536, 0.03993, 0.07677, 0.01388, 0.0029, 0.00235, 0.05823, 0.05237, 0.00425, 0.09225, 0.00703, 0.24038, 0.06733, 0.00064, 0.08959, 0.04365, 0.02308, 0.04566, 0.08395, 0.0038, 0.05322, 0.0145, 0.02012, 0.07084, 0.08202, 0.01091, 0.03738, 0.03798, 0.03473, 0.08534, 0.00133, 0.04046, 0.10119, 0.0317, 0.00312, 0.03614, 0.10442, 0.13286, 0.0042, 0.04229, 0.01735, 0.09879, 0.07516, 0.00303, 0.08062, 0.09347, 0.03473, 0.05099, 0.16373, 0.08988, 0.04696, 0.07488, 0.12159, 0.11098, 0.00549, 0.00122, 0.05276, 0.09883, 0.01346, 0.02059, 0.07394, 0.0413, 0.08766, 0.0124, 0.09913, 0.00754, 0.15671, 0.02699, 0.09978, 1e-05, 0.00243, 0.02819, 0.00027, 0.05793, 0.03165, 0.10168, 0.00042, 0.00044, 0.01332, 0.00542, 0.05946, 0.009, 0.10857, 0.01699, 1e-05, 0.00073, 0.10842, 0.17143, 0.00036, 0.00014, 0.10508, 0.01333, 0.34202, 0.12201, 0.04618, 0.02507, 0.02939, 0.03497, 0.01905, 0.00136, 0.02354, 0.00061, 0.08514, 0.14529, 0.04097, 0.12821, 0.18862],
[0.04683, 0.02943, 0.07885, 0.07846, 0.06855, 0.02815, 0.00792, 0.0826, 0.00554, 0.01041, 0.03957, 0.0126, 0.08399, 0.15046, 0.15594, 0.03941, 0.0428, 0.11343, 0.15665, 0.07381, 0.04386, 0.12008, 0.04816, 0.04844, 0.08248, 0.08023, 0.03011, 0.00464, 0.07204, 0.08376, 0.05777, 0.06164, 0.00697, 0.02023, 0.04844, 0.0592, 0.00954, 0.06357, 0.0122, 0.05905, 0.00705, 0.0054, 0.08822, 0.06056, 0.02598, 0.02136, 0.05638, 0.03768, 0.05101, 0.08908, 0.0384, 0.01579, 0.04023, 0.03746, 0.17236, 0.08293, 0.12469, 0.14018, 0.04301, 0.07258, 0.02678, 0.08078, 0.07698, 0.06346, 0.06984, 0.04832, 0.07512, 0.0342, 0.05339, 0.026, 0.11585, 0.02744, 0.00979, 0.01312, 0.05915, 0.01326, 0.00107, 0.00737, 0.05971, 0.0451, 0.05788, 0.0007, 0.0043, 0.00142, 0.0019, 0.00055, 0.00223, 0.02441, 0.04555, 0.03869, 0.05791, 0.05517, 0.15743, 0.04517, 0.47114, 0.05639, 0.00152, 0.00371, 1e-05, 1e-05, 0.04192, 0.02758, 0.01945, 0.02763, 0.04021, 0.02844, 0.01823, 0.10665, 0.02067, 0.05433, 0.05591, 0.00733, 0.00858, 0.01949, 0.06519, 0.07793, 0.00199, 0.09916, 0.08717, 0.06273, 0.09408, 0.00638, 0.00248, 0.08922, 0.09157, 0.03525, 0.01791, 0.06016, 0.01939, 0.12194, 0.08303, 0.0831, 0.02714, 0.06312, 0.11584, 0.11334, 0.04314, 0.02575, 0.00629, 0.02408, 0.02274, 0.03037, 0.06737, 0.0175, 0.00888, 0.06568, 0.0839, 0.0085, 0.00831, 0.00154, 0.01072, 0.01289, 0.09074, 0.02131, 0.02997, 0.02343, 0.02355, 0.05324, 0.09564, 0.17995, 0.00828, 0.0148, 0.01858, 0.02106, 0.00288, 0.00344, 0.001, 0.02143, 0.00732, 0.01458, 0.01547, 0.01742, 0.00032, 0.24005, 0.00028, 0.00302, 0.07275, 0.04579, 0.06316, 0.02572, 0.09316, 0.03062, 0.10521, 0.07123, 0.03069, 0.07958, 0.04484, 0.01948, 0.01951, 0.01282, 0.00868, 0.07931, 0.01105, 0.01235, 0.09297, 0.06959, 0.00716, 0.0271, 0.00592, 0.09362, 0.00319, 0.00859, 0.08486, 0.02001, 0.00194, 0.04189, 0.09024, 0.07705, 0.07365, 0.01123, 0.03202, 0.01361, 0.00098, 0.00397, 0.00139, 0.00397, 0.00445, 1e-05, 0.00267, 0.06564, 0.06567, 0.06566, 0.06566, 0.09249, 0.03475, 0.0338, 0.0664, 0.02986, 0.04024, 0.00835, 0.04304, 0.04081, 0.04534, 0.06636, 0.03312, 0.06175, 0.03117, 0.02243, 0.03454, 0.11135, 0.07016, 0.0681, 0.09716, 0.02589, 0.4367, 0.08293, 0.11834, 0.00191, 0.10913, 0.00159, 0.0638, 0.01808, 0.00116, 0.00911, 0.01408, 0.09179, 0.02122, 0.05026, 0.05144, 0.03169, 0.06674]]

log_data = [[np.log10(d) for d in row] for row in data]

fig, ax = plt.subplots(1, 3, sharey=True)
sns.violinplot(data=log_data, ax=ax[0])
sns.swarmplot(data=log_data, ax=ax[1])
sns.stripplot(data=log_data, ax=ax[2])
ax[0].yaxis.set_major_formatter(mticker.StrMethodFormatter("$10^{{{x:.0f}}}$"))
ax[0].yaxis.set_ticks([np.log10(x) for p in range(-6,1) for x in np.linspace(10**p, 10**(p+1), 10)], minor=True)
plt.show()

这应该显示预期的情节.

This should show the expected plot.

这篇关于Python 中的 Violin Plot 在对数尺度上的问题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！