pandas 有没有办法根据不同列中的离散变量制作 matplotlib 散点图标记或颜色？
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/24297097/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
Is there a way to make matplotlib scatter plot marker or color according to a discrete variable in a different column?
提问by Charlie_M
I'm making scatterplots out of a DF using matplotlib. In order to get different colors for each data set, I'm making two separate calls to plt.scatter:
我正在使用 matplotlib 从 DF 中制作散点图。为了为每个数据集获得不同的颜色，我对 plt.scatter 进行了两次单独的调用：
plt.scatter(zzz['HFmV'], zzz['LFmV'], label = dut_groups[0], color = 'r' )
plt.scatter(qqq['HFmV'], qqq['LFmV'], label = dut_groups[1], color = 'b' )
plt.legend()
plt.show()
This gives me the desired color dependence but really what would be ideal is if I could just get pandas to give me the scatterplot with several datasets on the same plot by something like
这给了我所需的颜色依赖性，但真正理想的是，如果我可以让Pandas通过类似的方式在同一图上为我提供带有多个数据集的散点图
df.plot(kind = scatter(x,y, color = df.Group, marker = df.Head)
df.plot(kind = scatter(x,y, color = df.Group, 标记 = df.Head)
Apparently there is no such animal (at least that I could find). So, next best thing in my mind would be to put the plt.scatter calls into a loop where I could make the color or marker vary according to one of the rows (not x or y, but some other row. If the row I want to use were a continuous variable it looks like I could use a colormap, but in my case the row I need to sue for this is a string ( categorical type of variable, not a number).
显然没有这样的动物（至少我能找到）。因此，我认为下一个最好的事情是将 plt.scatter 调用放入一个循环中，我可以在其中根据行之一（不是 x 或 y，而是其他行）使颜色或标记发生变化。如果行我想要使用的是一个连续变量，看起来我可以使用颜色图，但在我的情况下，我需要为此起诉的行是一个字符串（变量的分类类型，而不是数字）。
Any help much appreciated.
非常感谢任何帮助。
回答by mwaskom
What you're doing will almost work, but you have to pass colora vector of colors, not just a vector of variables. So you could do:
你正在做的几乎可以工作，但你必须传递color一个颜色向量，而不仅仅是一个变量向量。所以你可以这样做：
color = df.Group.map({dut_groups[0]: "r", dut_groups[1]: "b"})
plt.scatter(x, y, color=color)
Same goes for the marker style
标记样式也是如此
You could also use seabornto do the color-mapping the way you expect (as discussed here), although it doesn't do marker style mapping:
你也可以使用seaborn做色彩映射你希望（如讨论的方式在这里），虽然它没有做标记样式映射：
import seaborn as sns
import pandas as pd
from numpy.random import randn
data = pd.DataFrame(dict(x=randn(40), y=randn(40), g=["a", "b"] * 20))
sns.lmplot("x", "y", hue="g", data=data, fit_reg=False)
