pandas 如何在条形图的条形内显示值？
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/44491583/
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
How to show values inside the bars of a bargraph?
提问by user517696
I have a dataframe like this:
我有一个这样的数据框：
platform count
release_year
1996 PlayStation 138
1997 PlayStation 170
1998 PlayStation 155
1999 PC 243...
Now I want to plot horizontal bargraph with the Platform name inside the respective bars such that it looks something like this:
现在我想在各自的条形内绘制带有平台名称的水平条形图，使其看起来像这样：
How do I do that?
我怎么做？
回答by David Jaimes
Here's the input data.csvfile once you find the percentage in each platform:
这是data.csv在每个平台中找到百分比后的输入文件：
Platform,Percent
Nintendo,34
PC,16
Playstation,28
Xbox,22
This is the code:
这是代码：
import pandas as pd
import matplotlib.pyplot as plt
df = pd.read_csv("data.csv", index_col=0)
df.plot(kind="barh", legend=False, width=0.8)
for i, (p, pr) in enumerate(zip(df.index, df["Percent"])):
plt.text(s=p, x=1, y=i, color="w", verticalalignment="center", size=18)
plt.text(s=str(pr)+"%", x=pr-5, y=i, color="w",
verticalalignment="center", horizontalalignment="left", size=18)
plt.axis("off")
# xticks & yticks have empty lists to reduce white space in plot
plt.xticks([])
plt.yticks([])
plt.tight_layout()
plt.savefig("data.png")
回答by void
Not sure if you want it to be in Percentage %or as count numberitself. That's up to you to decide. However first convert your dataframe into a list using:
不确定您是希望它是百分比还是计数本身。这由你来决定。但是首先使用以下方法将您的数据框转换为列表：
count = df["count"].tolist()
platform = df["platform"].tolist()
I will not be focusing on that. You can find some help regarding that from
我不会专注于此。你可以从
Once you get the below list then,
一旦你得到下面的清单，
count = ['138','170','155','243','232']
platform =['PlayStation','PlayStation','PlayStation','PC','PlayStation']
Note: The above two would be your text labelsinside bar graphs.
注意：以上两个将是条形图中的文本标签。
Here is the complete code:
这是完整的代码：
import matplotlib.pyplot as plt
from numpy.random import rand
from numpy import arange
count = ['138','170','155','243','232']
platform =['PlayStation','PlayStation','PlayStation','PC','PlayStation']
def autolabel(rects):
# attach some text labels
for ii,rect in enumerate(rects):
width = int(rect.get_width())
height = rect.get_height()
print(height,width)
yloc1=rect.get_y() + height /2.0
yloc2=rect.get_y() + height /2.0
if (width <= 5):
# Shift the text to the right side of the right edge
xloc1 = width + 1
yloc2=yloc2+0.3
# Black against white background
clr = 'black'
align = 'left'
else:
# Shift the text to the left side of the right edge
xloc1 = 0.98*width
# White on blue
clr = 'white'
align = 'right'
yloc1=rect.get_y() + height /2.0
print(xloc1,yloc1,yloc2)
ax.text(xloc1,yloc1, '%s'% (count[ii]),horizontalalignment=align,
verticalalignment='center',color=clr,weight='bold',
clip_on=True)
ax.text(5,yloc2, '%s'% (platform[ii]),horizontalalignment='left',
verticalalignment='center',color=clr,weight='bold',
clip_on=True)
val = [138,170,155,243,232]
print(val)# the bar lengths or count in your case
pos = [ 1996 , 1997, 1998, 1999, 2000] # the bar centers on the y axis
print(pos)
fig = plt.figure()
ax = fig.add_subplot(111)
rects = ax.barh(pos,val, align='center',height=0.4)
print(rects)
autolabel(rects)
ax.set_ylabel('Year')
ax.set_xlabel('Count')
ax.set_title('horizontal bar chart')
ax.grid(False)
plt.savefig("horizontal.png")
plt.show()
The part where you will be interestedin very much:
你会非常感兴趣的部分：
def autolabel(rects):
# attach some text labels
for ii,rect in enumerate(rects):
width = rect.get_width()
height = rect.get_height()
yloc1=rect.get_y() + height /2.0
yloc2=rect.get_y() + height /2.0
if (width <= 5):
# Shift the text to the right side of the right edge
xloc1 = width + 1
yloc2=yloc2+0.3
# Black against white background
clr = 'black'
align = 'left'
else:
# Shift the text to the left side of the right edge
xloc1 = 0.98*width
# White on blue
clr = 'white'
align = 'right'
yloc1=rect.get_y() + height /2.0
ax.text(xloc1,yloc1, '%s'% (count[ii]),horizontalalignment=align,
verticalalignment='center',color=clr,weight='bold',
clip_on=True)
ax.text(5,yloc2, '%s'% (platform[ii]),horizontalalignment='left',
verticalalignment='center',color=clr,weight='bold',
clip_on=True)
1) iivariable comes from enumerate having values 0 to 5. Used to iterate over our lists countand platform
1) ii变量来自枚举值 0 到 5。用于迭代我们的列表count和platform
2) Why an if/else statement in the function? That is for conditions where the width is too little. Say if the first width obtained from val = [138,170,155,243,232]is reduced to 5 i.e val = [5,170,155,243,232]in this case the output would be.
2) 为什么在函数中有一个 if/else 语句？那是针对宽度太小的情况。假设从 获得的第一个宽度val = [138,170,155,243,232]减少到 5，即val = [5,170,155,243,232]在这种情况下输出将是。
What we are basically doing is giving xloc (x-coordinate) and yloc (y-coordinate) values for both ax.text()functions.
我们基本上要做的是为这两个ax.text()函数提供 xloc（x 坐标）和 yloc（y 坐标）值。
ax.text(xloc1,yloc1, '%s'% (count[ii]),horizontalalignment=align,
verticalalignment='center',color=clr,weight='bold',
clip_on=True)
ax.text(5,yloc2, '%s'% (platform[ii]),horizontalalignment='left',
verticalalignment='center',color=clr,weight='bold',
clip_on=True)
Function parameters
text(x, y, s, fontdict=None, withdash=False, **kwargs)
x, y : data coordinates
s : string while the other two are optional.
功能参数
文本（x，y，s，fontdict=None，withdash=False，**kwargs）
x, y : 数据坐标
s : 字符串，而其他两个是可选的。
If width is < 5. Then increase yloc a bit. So that text is little bit higher. While change xloc acccordingly.Also changing the color to black. Or else color will be white.
如果宽度 < 5。然后稍微增加 yloc。所以那个文本要高一点。相应地更改 xloc。同时将颜色更改为黑色。否则颜色会是白色的。
It would be bestif you change those values and see how the output changes to gain a better understanding of it.
如果您更改这些值并查看输出如何变化以更好地理解它，那将是最好的。
UPDATE:If you don't want the axis to be shown in the output just as in the image you attached you can simple do that typing in ax.axis("off")
更新：如果您不希望轴显示在输出中，就像您附加的图像一样，您可以简单地输入ax.axis("off")
