数据可视化已成为与分析数据进行交流的重要阶段。通过数据可视化，数据科学家和业务分析师可以轻松地从大量数据中提取见解。
Seaborn 是一种 Python 中的统计图形绘图和可视化库，允许数据分析师和数据科学专业人员呈现可视化。
在本文中，我们将讨论如何通过强大的seaborn库创建和处理条形图。
条形图或条形图是最突出的可视化图之一，用于以图形格式描述统计和数据驱动的操作。
统计学家和工程师用它来显示数字变量和分类变量之间的关系。
分类变量的所有实体都以条形的形式表示。条形大小代表数值。
我们可以使用Pandas 的数据框用于可视化seaborn中的数据。为此，我们必须导入三个模块：Pandas、Matplotlib和海生。这是显示如何使用它的代码片段。
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
#creating dataframe
df = pd.DataFrame({
'Name':list("QWERTYUI"),
'Student IDs': list(range(12001,12009)),
'Yearly Progress':list('ABCDABCD'),
'Chemistry': [59,54,42,66,60,78,64,82],
'Physics': [90,42,88,48,80,73,43,69],
'Math': [45,85,54,64,72,64,67,70]
})
df_detailed = pd.melt(df, id_vars=['Name', 'Student IDs', 'Yearly Progress'], value_vars=['Chemistry', 'Physics', 'Math'])
ax = sns.barplot(x = "Yearly Progress", y = "value", hue = "variable", data = df_detailed)
plt.show() Output
我们还可以使用一个充满数据的简单列表来使用 seaborn 生成条形图，而不是使用 Pandas DataFrame。这是一个代码片段，展示了如何使用它。
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
raw_data1 = [[127.66773437098896, 47.585408826566024, 90.426437828641106, 414.955787935926836],
[147.700582993718115, 37.59796443553682, 245.38896827262796, 312.80916093973529],
[123.66563311651776, 177.476571906259835, 145.21460968763448, 54.78683755963528],
[47.248523637295705, 67.42573841363118, 145.52890109500238, 211.10243082784969],
[148.14532745960979, 87.46958795222966, 67.4804195003332, 68.97715435208194],
[46.61620129160194, 224.316775886868584, 138.053032014046366, 41.527497508033704]]
def plot_sns(raw_data):
data = np.array(raw_data)
x = np.arange(len(raw_data))
sns.axes_style('white')
sns.set_style('white')
ax = sns.barplot(x, data[:,0])
plt.show()
plot_sns(raw_data1) Output
在条形图上或条形内显示各个值是使用 seaborn 条形图的另一种重要方式。
我们可以使用带有一些属性的 annotate() 来设置标签发生的位置。下面是一个代码片段，展示了如何通过DataFrame整体实现它。
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
continent_data = ["Asia", "America", "Africa", "Oceania", "Europe"]
lifeExpectancy = [87, 63, 77, 56, 39]
dataf = pd.DataFrame({"continent_data":continent_data, "lifeExpectancy":lifeExpectancy})
plt.figure(figsize = (9, 7))
splot = sns.barplot(x = "continent_data", y = "lifeExpectancy", data=dataf)
for g in splot.patches:
splot.annotate(format(g.get_height(), '.1f'),
(g.get_x() + g.get_width() / 2., g.get_height()),
ha = 'center', va = 'center',
xytext = (0, 9),
textcoords = 'offset points')
plt.xlabel("Continent Names", size = 14)
plt.ylabel("Life Expectancy", size = 14)
plt.show() Output
We can also provide the value within the bars in the barplot. To do this we have to simply set the xytext attribute to (0, -12), which means the y value of the attribute will go downward because of the negative value set for it. In this case, the code will be:
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
continent_data = ["Asia", "America", "Africa", "Oceania", "Europe"]
lifeExpectancy = [87, 63, 77, 56, 39]
dataf = pd.DataFrame({"continent_data":continent_data, "lifeExpectancy":lifeExpectancy})
plt.figure(figsize = (9, 7))
splot = sns.barplot(x = "continent_data", y = "lifeExpectancy", data=dataf)
for g in splot.patches:
splot.annotate(format(g.get_height(), '.1f'),
(g.get_x() + g.get_width() / 2., g.get_height()),
ha = 'center', va = 'center',
xytext = (0, -13),
textcoords = 'offset points')
plt.xlabel("Continent Names", size = 14)
plt.ylabel("Life Expectancy", size = 14)
plt.show() Output
有多种方法可以更改默认箱线图的颜色。我们可以使用 barplot() 的 color 属性将它们全部更改为新颜色。
我们还可以通过列表手动设置调色板值或使用已经使用seaborn创建的默认调色板颜色集。以下代码片段展示了如何将所有条形更改为单一颜色。
import matplotlib.pyplot as plt
import seaborn as sns
x = ['A', 'B', 'C', 'D']
y = [1, 5, 3, 7]
sns.barplot(x, y, color = 'violet')
plt.show() Output
Again we can manually set colors for each bar using the palette parameter of the barplot(). We have to pass the color names within the list as palette values.
import matplotlib.pyplot as plt
import seaborn as sns
x = ['A', 'B', 'C', 'D']
y = [1, 5, 3, 7]
sns.barplot(x, y, color = 'blue', palette = ['tab:blue', 'tab:orange', 'tab:red', 'tab:green'])
plt.show() Output
Apart from manually providing the color values, we can also use predefined color palettes defined in the Seaborn color palettes. Here is how to use it.
import matplotlib.pyplot as plt
import seaborn as sns
x = ['A', 'B', 'C', 'D']
y = [1, 5, 3, 7]
sns.barplot(x, y, color = 'blue', palette = 'hls')
plt.show() Output
我们还可以对条形图值或数据设置特定条件，为它们设置不同的颜色或调色板。
以下代码片段展示了如何根据条件通过调色板设置不同的颜色。
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
continent_data = ["Asia", "America", "Africa", "Oceania", "Europe"]
lifeExpectancy = [99, 63, 77, 56, 27]
custom_palette = {}
dataf = pd.DataFrame({"continent_data":continent_data, "lifeExpectancy":lifeExpectancy})
plt.figure(figsize = (9, 7))
for q in lifeExpectancy:
if q < 30 and q > 50:
custom_palette[q] = 'Paired'
elif q < 50 and q > 60:
custom_palette[q] = 'PRGn'
elif q < 60 and q > 70:
custom_palette[q] = 'husl'
else:
custom_palette[q] = 'hls'
sns.barplot(x = "continent_data", y = "lifeExpectancy", data=dataf, palette=custom_palette[q])
plt.xlabel("Continent Names", size = 14)
plt.ylabel("Life Expectancy", size = 14)
plt.show() Output
我们可以使用 barplot() 生成水平条形图。它还将具有相同的功能和可视化能力；唯一的区别是，它们不是垂直放置（我们到目前为止已经遇到过），而是水平放置。
import seaborn as sns
import matplotlib.pyplot as plt
labels = ['C++', 'Python', 'Go', 'Ruby', 'JavaScript']
price = [50, 60, 90, 100, 150]
sns.barplot(x = price, y = labels)
plt.show() Output
数据可视化中的图例是存在于图表任意一角的小方框。它包含与表示绘图的某些元素类型的文本关联的多条颜色线。
当多个数据驻留在图表中时，图例中的指示表示哪个组件代表哪个数据。
Seaborn 将从 DataFrame 中获取字符串（键）作为图例标签。下面是一段代码片段，展示了seaborn如何生成一个。
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
df = pd.DataFrame({'Scores': [24, 15, 17, 11, 16, 21, 24, 27],
'assists': [1, 8, 9, 12, 11, 6, 3, 1],
'Group': ['A', 'A', 'A', 'A', 'B', 'B', 'B', 'B']})
sns.barplot(data = df, x = 'Scores', y = 'assists', hue = 'Group')
plt.legend(title = 'Group Category')
plt.show()
Output
We can also change the position of the Legend but not using seaborn. For such manipulation, we need to use the Matplotlib.pyplot’s legend() method loc parameter. We can also specify other location values to the loc:
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
df = pd.DataFrame({'Scores': [24, 15, 17, 11, 16, 21, 24, 27],
'assists': [1, 8, 9, 12, 11, 6, 3, 1],
'Group': ['A', 'A', 'A', 'A', 'B', 'B', 'B', 'B']})
sns.barplot(data = df, x = 'Scores', y = 'assists', hue = 'Group')
plt.legend(loc = 'center right', title = 'Group Category')
plt.show() Output
有两种不同的方法可以从seaborn 的可视化中删除图例。这些都是：
Method 1：使用 matplotlib.pyplot.legend()：我们可以使用 matplotlib.pyplot.legend() 函数向 seaborn 图中添加自定义图例。
我们可以使用 matplotlib.pyplot 中的 legend()，因为 Seaborn 运行在 matplotlib 模块之上，因此很容易操纵 Seaborn 操作和可视化。
要删除图例，我们必须向图中添加一个空图例并删除其框架。它将有助于隐藏最终图形中的图例。
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
df = pd.DataFrame({'Scores': [24, 15, 17, 11, 16, 21, 24, 27],
'assists': [1, 8, 9, 12, 11, 6, 3, 1],
'Group': ['A', 'A', 'A', 'A', 'B', 'B', 'B', 'B']})
sns.barplot(data = df, x = 'Scores', y = 'assists', hue = 'Group')
plt.legend([],[], frameon=False)
plt.show() Output
Method 2：使用remove()：我们还可以使用legend_.remove()从图中删除图例。
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
df = pd.DataFrame({'Scores': [24, 15, 17, 11, 16, 21, 24, 27],
'assists': [1, 8, 9, 12, 11, 6, 3, 1],
'Group': ['A', 'A', 'A', 'A', 'B', 'B', 'B', 'B']})
gk = sns.barplot(data = df, x = 'Scores', y = 'assists', hue = 'Group')
gk.legend_.remove()
plt.show() Output
我们经常需要标记 x 轴和 y 轴，以便更好地指示或赋予绘图含义。要在图中设置标签，有两种不同的方法。这些都是：
Method 1：使用 set() 方法：第一种方法是使用 set() 方法并将字符串作为标签传递给 xlabel 和 ylabel 参数。这是一个代码片段，展示了我们如何执行此操作。
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
datf = pd.DataFrame({"Season 1": [7, 4, 5, 6, 3],
"Season 2" : [1, 2, 8, 4, 9]})
p = sns.barplot(data = datf)
p.set(xlabel="X Label Value", ylabel = "Y Label Value")
plt.show() Output
Method 2: Using matplotlib’s xlabel() and ylabel(): Since seaborn runs on top of Matplotlib, we can use the Matplotlib methods to set the labels for the x and y axes. Here is a code snippet showing how can we perform that.
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
datf = pd.DataFrame({"Season 1": [7, 4, 5, 6, 3],
"Season 2" : [1, 2, 8, 4, 9]})
p = sns.barplot(data = datf)
plt.xlabel('X axis labeling')
plt.ylabel('Y axis labeling')
plt.show() Output
默认情况下，seaborn 中的条形图不提供显式参数来更改条形宽度。但我们可以编写并添加一个单独的函数来执行此操作。
import pylab as plt
import seaborn as sns
tips = sns.load_dataset("tips")
fig, axi = plt.subplots()
sns.barplot(data = tips, ax = axi, x = "time", y = "tip", hue = "sex")
def width_changer(axi, new_value):
for patch in axi.patches :
cur_width = patch.get_width()
diff = cur_width - new_val
patch.set_width(new_val)
patch.set_x(patch.get_x() + diff * .5)
plt.legend(loc = 'upper right', title = 'Gender')
width_changer(axi, .18)
plt.show()
Output
在seaborn中设置绘图的图形大小有不同的方法。一些众所周知的方法是：
Method 1:seaborn.set():
此方法有助于控制seaborn 图的组成和配置。我们可以设置rc参数并传递图形大小属性来设置seaborn图的图形大小。
这是显示如何使用它的代码片段。
import pylab as plt
import seaborn as sns
tips = sns.load_dataset("tips")
fig, axi = plt.subplots()
sns.set(rc = {'figure.figsize':(12.0,8.5)})
sns.barplot(data = tips, ax = axi, x = "time", y = "tip", hue = "sex")
def width_changer(axi, new_val):
for patch in axi.patches :
cur_width = patch.get_width()
diff = cur_width - new_val
patch.set_width(new_val)
patch.set_x(patch.get_x() + diff * .5)
plt.legend(loc = 'upper right', title = 'Gender')
width_changer(axi, .18)
plt.show() Output
Method 2: matplotlib.pyplot.figure(): We can also use the matplotlib.pyplot.figure() and use the figsize parameter to pass two values as Tuple. The code snippet will look something like this.
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from matplotlib import rcParams
datf = pd.DataFrame({"Season 1": [7, 4, 5, 6, 3],
"Season 2" : [1, 2, 8, 4, 9]})
plt.figure(figsize = (14, 7))
p = sns.barplot(data = datf)
plt.legend(loc = 'center', title = 'Karlos TV series')
plt.show() Output
Method 3: Using rcParams: We can also use the rcParams, which is a part of the Matplotlib library to control the style and size of the plot. It works similar to the seaborn.set(). Here is a code snippet showing how to use it.
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from matplotlib import rcParams
datf = pd.DataFrame({"Season 1": [7, 4, 5, 6, 3],
"Season 2" : [1, 2, 8, 4, 9]})
rcParams['figure.figsize'] = 14, 7
p = sns.barplot(data = datf)
plt.legend(loc = 'center', title = 'Karlos TV series')
plt.show() Output
Method 4: matplotlib.pyplot.gcf(): This function helps in getting the instance of the current figure. Along with the gcf() instance, we can use the set_size_inches() method to modify the final size of the seaborn plot. The code snippet will look something like this.
import pylab as plt
import seaborn as sns
tips = sns.load_dataset("tips")
fig, axi = plt.subplots()
sns.barplot(data = tips, ax = axi, x = "time", y = "tip", hue = "sex")
def width_changer(axi, new_val):
for patch in axi.patches :
cur_width = patch.get_width()
diff = cur_width - new_val
patch.set_width(new_val)
patch.set_x(patch.get_x() + diff * .5)
plt.gcf().set_size_inches(14, 7)
plt.legend(loc = 'upper right', title = 'Gender')
width_changer(axi, .18)
plt.show() Output
在通过seaborn 创建可视化效果时，字体大小起着重要作用。有两种不同的方法可以设置可视化的字体大小。这些都是：
Method 1：使用 fontsize 参数：我们可以将此参数与多个 matplotlib 方法一起使用，例如 xlabel()、ylabel()、title() 等。下面是一个代码片段，展示如何利用它们。
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
datf = pd.DataFrame({'dor': ['1/12/2022', '1/30/2022', '2/27/2022', '2/28/2022'],
'marketReach': [7, 12, 5, 16],
'Brand': ['A', 'A', 'B', 'B']})
sns.barplot(x = 'dor', y = 'marketReach', hue = 'Brand', data = datf)
plt.legend(title = 'Brand Name', fontsize=20)
plt.xlabel('Date of Release', fontsize = 15);
plt.ylabel('Market Reach', fontsize = 15);
plt.title('Overall Sales Report', fontsize = 22)
plt.tick_params(axis='both', which='major', labelsize=12)
plt.show() Output
Method 2: Using the set() method: Another way to set the font size for all the fonts associated with the plot is using the set() method of the Seaborn. Here’s how to use it.
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
datf = pd.DataFrame({'dor': ['1/12/2022', '1/30/2022', '2/27/2022', '2/28/2022'],
'marketReach': [7, 12, 5, 16],
'Brand': ['A', 'A', 'B', 'B']})
sns.set(font_scale = 3)
sns.barplot(x = 'dor', y = 'marketReach', hue = 'Brand', data = datf)
plt.legend(title = 'Brand Name', fontsize=20)
plt.show() Output
我们可以通过两种不同的方法在酒吧之间创造空间。
Method 1：通过在连续条之间添加空白条：
import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd
datf = pd.DataFrame({'Name': ['Karl', 'Ray', 'Sue', 'Dee'],
'SalInLac': [25, 28, 21, 26], 'Gender': ['M', 'M', 'F', 'F']})
datf = pd.concat([datf[datf.Gender == 'M'],
pd.DataFrame({'Name': [''], 'SalInLac': [0], 'Gender': ['M']}), datf[datf.Gender == 'F']])
age_plot = sns.barplot(data = datf, x = "Name", y = "SalInLac", hue="Gender", dodge=False)
plt.setp(age_plot.get_xticklabels(), rotation=90)
plt.ylim(0, 45)
age_plot.tick_params(labelsize = 6)
age_plot.tick_params(length = 4, axis='x')
age_plot.set_ylabel("Age", fontsize=12)
age_plot.set_xlabel("", fontsize=1.5)
plt.tight_layout()
plt.show() Output
Method 2: By creating spaces in the label names:
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
datf = pd.DataFrame({'val': [' Val1 ', ' Val2 ', ' Val3 ', ' Val4 '],
'marketReach': [7, 12, 5, 16],
'Brand': ['A', 'A', 'B', 'B']})
sns.set(font_scale = 1)
sns.barplot(x = 'val', y = 'marketReach', hue = 'Brand', data = datf)
plt.legend(title = 'Brand Name', fontsize=20)
plt.show() Output
通过调整刻度标签有两种不同的旋转轴的方法。
Method 1：使用 xticks() 方法：matplotlib.pyplot.xticks() 函数中可用的旋转参数可以帮助实现此目的。这是有关如何使用它的代码片段。
import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd
datf = pd.DataFrame({'Name': ['Karl', 'Ray', 'Sue', 'Dee'], 'SalInLac': [25, 28, 21, 26], 'Gender': ['M', 'M', 'F', 'F']})
datf = pd.concat([datf[datf.Gender == 'M'], pd.DataFrame({'Name': [''], 'SalInLac': [0], 'Gender': ['M']}), datf[datf.Gender == 'F']])
age_plot = sns.barplot(data = datf, x = "Name", y = "SalInLac", hue="Gender", dodge=False)
plt.setp(age_plot.get_xticklabels(), rotation=90)
plt.ylim(0, 45)
age_plot.tick_params(labelsize = 10)
age_plot.tick_params(length = 4, axis='x')
age_plot.set_ylabel("Age", fontsize=12)
age_plot.set_xlabel("", fontsize=1.5)
plt.tight_layout()
plt.xticks(rotation = 45)
plt.show() Output
Method 2: using setp() method: The set parameter setp() method also enables us to rotate the x-axis tick-levels. Here is the code snippet on how to use it.
import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd
datf = pd.DataFrame({'Name': ['Karl', 'Ray', 'Sue', 'Dee'], 'SalInLac': [25, 28, 21, 26], 'Gender': ['M', 'M', 'F', 'F']})
datf = pd.concat([datf[datf.Gender == 'M'], pd.DataFrame({'Name': [''], 'SalInLac': [0], 'Gender': ['M']}), datf[datf.Gender == 'F']])
age_plot = sns.barplot(data = datf, x = "Name", y = "SalInLac", hue="Gender", dodge=False)
plt.setp(age_plot.get_xticklabels(), rotation=90)
plt.ylim(0, 45)
age_plot.tick_params(labelsize = 10)
age_plot.tick_params(length = 4, axis='x')
age_plot.set_ylabel("Age", fontsize=12)
age_plot.set_xlabel("", fontsize=1.5)
plt.tight_layout()
locs, labels = plt.xticks()
plt.setp(labels, rotation = 45)
plt.show() Output
我们可以在 Seaborn 中对多个条形图进行分组，并将它们显示在一个图下。这是显示如何使用它的代码片段。
import seaborn as sns
import matplotlib.pyplot as plt
tips = sns.load_dataset("tips")
colors = ["#69d", "#69b3a2"]
sns.set_palette(sns.color_palette(colors))
plt.figure(figsize = (9, 8))
# multiple barplot grouped together
ax = sns.barplot(
x = "day",
y = "total_bill",
hue = "smoker",
data = tips,
ci=None
)
ax.set_title("Smokers with large bills")
ax.set_ylabel("Expense")
plt.show() Output
我们可以使用 sort_values() 方法对 Seaborn 条形图中的条形进行排序，并传递排序所依据的对象。
另外，我们必须将 ascending 参数设置为 True 或 False 来确定是否按升序或降序显示排序条形图。
下面是一个代码片段，展示了如何在 seaborn.barplot() 中使用它。
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
CompanyLocations = ["Delhi", "Mumbai", "Madhya Pradesh",
"Hyderabad", "Bangaluru"]
CompanyProfit = [644339, 780163, 435245, 256079, 805463]
datf = pd.DataFrame({"CompanyLocations": CompanyLocations,
"CompanyProfit": CompanyProfit})
# make barplot and sort bars
sns.barplot(x='CompanyLocations',
y="CompanyProfit", data = datf,
order = datf.sort_values('CompanyProfit',ascending = False).CompanyLocations)
plt.show() Output
If we want to show it in ascending order, we have to simply set the ascending = True
sns.barplot(x='CompanyLocations',
y="CompanyProfit", data = datf,
order = datf.sort_values('CompanyProfit',ascending = True).CompanyLocations) Output
我们希望本文能够全面介绍什么是seaborn条形图以及如何使用seaborn创建条形图、创建条件、包含和排除图例、设置图形大小、字体和标签。
我们还收集了有关如何更改与 barplot() 相关的各种配置的见解和编程经验。
由于seaborn barplot在数据可视化和数据分析绘制图表中非常重要，因此我们应该充分理解它。