Pandas Seaborn Swarmplot 不绘图

Question

我正在尝试绘制一个 seaborn swarmplot，其中 col[2] 是频率， col[3] 是要分组的类。下面给出了输入和代码。 输入

tweetcricscore,51,high active
tweetcricscore,46,event based
tweetcricscore,12,event based
tweetcricscore,46,event based
tweetcricscore,1,viewers 
tweetcricscore,178,viewers
tweetcricscore,46,situational
tweetcricscore,23,situational
tweetcricscore,1,situational
tweetcricscore,8,situational
tweetcricscore,56,situational

代码：

import numpy as np
import pandas as pd
import matplotlib as mpl
import matplotlib.pyplot as plt
import seaborn as sns

sns.set(style="whitegrid", color_codes=True)

df = pd.read_csv('input.csv', header = None)

df.columns = ['keyword','freq','class']

ax = sns.swarmplot(x="class", y="freq", data=df)

plt.show()

代码不会绘图，也不会给出任何错误。有什么优化代码的建议吗？

Answer 1

我认为您首先需要read_csv，然后通过连接fillna 和最后一个strip 空格来创建新的列类：

import pandas as pd
import matplotlib as mpl
import matplotlib.pyplot as plt
import seaborn as sns
import io

temp=u"""tweetcricscore 51 high active
tweetcricscore 46 event based
tweetcricscore 12 event based
tweetcricscore 46 event based
tweetcricscore 1 viewers 
tweetcricscore 178 viewers
tweetcricscore 46 situational
tweetcricscore 23 situational
tweetcricscore 1 situational
tweetcricscore 8 situational
tweetcricscore 56 situational"""
#after testing replace io.StringIO(temp) to filename
df = pd.read_csv(io.StringIO(temp), 
                 sep="\s+", #separator is arbitrary whitespace
                 names=['keyword','freq','class1','class2']) #set new col names

df['class'] = df['class1'] + ' ' + df['class2'].fillna('')
df['class'] = df['class'].str.strip()
print df
           keyword  freq       class1  class2        class
0   tweetcricscore    51         high  active  high active
1   tweetcricscore    46        event   based  event based
2   tweetcricscore    12        event   based  event based
3   tweetcricscore    46        event   based  event based
4   tweetcricscore     1      viewers     NaN      viewers
5   tweetcricscore   178      viewers     NaN      viewers
6   tweetcricscore    46  situational     NaN  situational
7   tweetcricscore    23  situational     NaN  situational
8   tweetcricscore     1  situational     NaN  situational
9   tweetcricscore     8  situational     NaN  situational
10  tweetcricscore    56  situational     NaN  situational

sns.set(style="whitegrid", color_codes=True)
ax = sns.swarmplot(x="class", y="freq", data=df)
plt.show()

class 列不包含空格时的解决方案：

import pandas as pd
import matplotlib as mpl
import matplotlib.pyplot as plt
import seaborn as sns
import io

temp=u"""tweetcricscore 51 highactive
tweetcricscore 46 eventbased
tweetcricscore 12 eventbased
tweetcricscore 46 eventbased
tweetcricscore 1 viewers 
tweetcricscore 178 viewers
tweetcricscore 46 situational
tweetcricscore 23 situational
tweetcricscore 1 situational
tweetcricscore 8 situational
tweetcricscore 56 situational"""
#after testing replace io.StringIO(temp) to filename
df = pd.read_csv(io.StringIO(temp), 
                 sep="\s+", #separator is arbitrary whitespace
                 names=['keyword','freq','class']) #set new col names
print df

           keyword  freq        class
0   tweetcricscore    51   highactive
1   tweetcricscore    46   eventbased
2   tweetcricscore    12   eventbased
3   tweetcricscore    46   eventbased
4   tweetcricscore     1      viewers
5   tweetcricscore   178      viewers
6   tweetcricscore    46  situational
7   tweetcricscore    23  situational
8   tweetcricscore     1  situational
9   tweetcricscore     8  situational
10  tweetcricscore    56  situational

sns.set(style="whitegrid", color_codes=True)
ax = sns.swarmplot(x="class", y="freq", data=df)
plt.show()

EDIT2：

如果分隔符是,，则使用：

import pandas as pd
import matplotlib as mpl
import matplotlib.pyplot as plt
import seaborn as sns
import io

temp=u"""tweetcricscore,51,high active
tweetcricscore,46,event based
tweetcricscore,12,event based
tweetcricscore,46,event based
tweetcricscore,1,viewers
tweetcricscore,178,viewers
tweetcricscore,46,situational
tweetcricscore,23,situational
tweetcricscore,1,situational
tweetcricscore,8,situational
tweetcricscore,56,situational"""
#after testing replace io.StringIO(temp) to filename
df = pd.read_csv(io.StringIO(temp), names=['keyword','freq','class'])

print df
           keyword  freq        class
0   tweetcricscore    51  high active
1   tweetcricscore    46  event based
2   tweetcricscore    12  event based
3   tweetcricscore    46  event based
4   tweetcricscore     1      viewers
5   tweetcricscore   178      viewers
6   tweetcricscore    46  situational
7   tweetcricscore    23  situational
8   tweetcricscore     1  situational
9   tweetcricscore     8  situational
10  tweetcricscore    56  situational

sns.set(style="whitegrid", color_codes=True)
ax = sns.swarmplot(x="class", y="freq", data=df)
plt.show()

Answer 2

在用超过8-10k 行的数据集绘制swamplot 并在jezreal 的不断帮助和建议下进行了几次尝试。我们得出的结论是seaborn 类别绘图swarmplot 无法像教程文档中提到的seaborn 中的其他绘图那样缩放大数据。因此，我将绘图样式更改为bokeh 散点图，其中我使用y 轴上的数值和x 轴上的分组类别名称，这有点解决了我用类别绘制univariate 数据的问题。

import numpy as np
import matplotlib.pyplot as plt
from pylab import*
import math
from matplotlib.ticker import LogLocator
import pandas as pd

from bokeh.models import BoxSelectTool, BoxZoomTool, LassoSelectTool
from bokeh.charts import Scatter, output_file, show
from bokeh.plotting import figure, hplot, vplot
from bokeh.models import LinearAxis

df = pd.read_csv('input.csv', header = None)

df.columns = ['user','freq','class']

scatter = Scatter( df, x='class', y='freq', color='class', marker='class', title=' User classification', legend=False)

output_file('output.html', title='output')

show(scatter)

这允许按class 列分组，并根据组分配颜色和标记。 freq 沿 y 轴绘制。

注意：这可能会意外地起作用，因为数据是离散的。