Seaborn的barplot bar和error bar的美学 [英] Aesthetics of barplot bars and error bars in seaborn

问题描述

我正在使用seaborn库来可视化数据，并且我想更改有关输出图的某些内容以供发布.我希望误差条更窄，有大写字母，我希望所有误差条周围的边框都是黑色的.

I'm using the seaborn library to visualize data and I want to change some things about the output graphs for publication. I want the error bars to be more narrow, with caps, and I'd like the border around all of the bars to be black.

我想象有某种方法可以使用pyplot(或sns.set_context中的rc词典)更改绘图，但是我不知道该怎么做.

I imagine that there is some way to change the plot using pyplot (or perhaps the rc dictionary in sns.set_context), but I can't figure out how to do this.

colors = ["black", "grey", "white"]
g = sns.barplot("TYPEMOD", "SCORE", ci=68, data=final_data,  palette=sns.xkcd_palette(colors))

我尝试过:

g.errorbar(capthick=2)

但这会产生错误，因为需要我忽略sns.barplot中的错误栏(这些错误栏是通过设置ci参数生成的，因此我将ci = None设置为零，然后执行全新的错误带有g.errorbar的条).我觉得必须有某种方法可以完成所有这些工作，因为这似乎是一个很小的更改，但是我无法在任何常规文档中找到任何内容.

But that gives an error, because requires me to just ignore the error bars in sns.barplot (which are generated by setting the ci parameter, so I'd set ci=None, and then I'd do completely new error bars with g.errorbar). I feel like there must be some way to do this without all of this effort, since it seems like a minor change, but I can't find anything in the seaborn documentation.

我还想将我的情节图中所有条周围的边框更改为黑色.

I'd also like to change the border around all of the bars in my barplot to be black.

推荐答案

barplot不是在内部使用errorbar，它只是在CI的间隔内画线，因此无法添加上限.误差线宽度本身只是lines.linewidth rc参数的比例因子，因此您可以临时设置它来控制它:

barplot isn't using errorbar under the hood, it's just drawing lines in the interval of the CI, so there's no way to add caps. The errorbar width itself is just a scaled factor of the lines.linewidth rc parameter, so you can set that temporarily to control it:

with mpl.rc_context("lines.linewidth": 1}):
    colors = ["black", "grey", "white"]
    g = sns.barplot("TYPEMOD", "SCORE", ci=68, data=final_data,
                    palette=sns.xkcd_palette(colors))

这篇关于Seaborn的barplot bar和error bar的美学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！