用matplotlib对数据可视化

下图是要用到的数据集，反映了从1984到2016年的失业率的变化

　1.导入可视化模块import matlibplot.pyplot as plt， 函数plt.plot(x, y)确定折线图的点,x是由这些点的x坐标组成的列表，y是由这些点的y坐标组成

的列表。plt.show()显示图像，plt.xlabel()给x轴命名，plt.xticks()可以设置x坐标刻度点旋转指定角度，plt.title()给折线图命名

　　下面的代码是以上函数的应用

 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 #画出1984年失业率折线图
 unrated = pd.read_csv("C:/学习/python/hello/UNRATE.csv")
 first_twelve = unrated.head()
 
 x_series = first_twelve["DATE"]
 y_series = first_twelve["VALUE"]
 
 plt.xticks(rotation=)
 plt.xlabel("Month")
 plt.ylabel("Unemployment rate")
 
 plt.title("Unemployment rate trends on 1984")
 
 plt.plot(x_series, y_series)
 plt.show()

　　运行结果如下

2.通过plt.subplot(n, m, x)在一个figure中添加多个子图, n和m表示子图的布局，分别代表行数和列数，x表示从左往右，从上往下数的第x个子图

下面的代码提供了该函数使用实例

 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 
 plt.figure(figsize=(, ))
 ax1 = plt.subplot(, , )
 ax2 = plt.subplot(, , )
 ax3 = plt.subplot(, , )
 ax5 = plt.subplot(, , )
 plt.show()

运行结果如下

3.下面的代码是在一个坐标轴中画多个折线图的示例

 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 
 unrated = pd.read_csv("C:/学习/python/hello/UNRATE.csv")
 unrated["DATE"] = pd.to_datetime(unrated["DATE"])
 color = ["red", "yellow", "blue", "green", "purple"]
 plt.figure(figsize=(, ))
 for i in range():
     sub_unrated = unrated.loc[i*:(i+)*-]
     sub_unrated_x = sub_unrated['DATE'].dt.month
     sub_unrated_y = sub_unrated["VALUE"]
     label = +i
     plt.plot(sub_unrated_x, sub_unrated_y, color=color[i], label=label)
 plt.legend(loc="best")
 plt.show()

运行结果如下

4.figure和subplot的定义顺序决定了subplot是画在哪个figure中。当代码中定义了多个figure时候，紧接着该figure定义的subplot才画在该figure中，

如下代码所示，定义了figure1和figure2，ax1和ax2在figure1中，ax在figure2中。

 plt.figure(figsize=(, ))
 ax1 = plt.subplot(,,)
 plt.figure(figsize=(, ))
 ax = plt.subplot(, , )
 
 plt.show()

5.用matplotlib画条形图

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
 
reviews = pd.read_csv("C:/学习/python/hello/fandango_score_comparison.csv")
cols = ["FILM", "RT_user_norm", "Metacritic_user_nom", "IMDB_norm", "Fandango_Ratingvalue", "Fandango_Stars"]
norm_reviews = reviews[cols]
 
num_cols = ["RT_user_norm", "Metacritic_user_nom", "IMDB_norm", "Fandango_Ratingvalue", "Fandango_Stars"]
bar_height = norm_reviews.loc[0, num_cols].values  #第一部电影的评价，注意利用loc索引某一行的用法，可以添加第二维
 
bar_position = 1 + np.arange(5)  #arange返回的是ndarray类型，range返回的是list类，使用arange需要用numpy
plt.figure(figsize=(10, 10))
ax = plt.subplot(1, 1, 1)
ax.bar(bar_position, bar_height, 0.3)  #bar_position是条形图的x坐标(中点坐标),bar_height是高，0.3是宽
#设置x坐标刻度
tick_positions = range(1, 6)
ax.set_xticks(tick_positions)
ax.set_xticklabels(num_cols)
 
#设置x轴和y轴名称
ax.set_xlabel("Rating Source")
ax.set_ylabel("Average Rating")
ax.set_title("Whatever")
plt.show()

运行结果如下

　　将上面的代码改变几处，就会成为横着的条形图了。代码如下所示（改动之处用白底红字加粗下划线标出来了）

 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 
 reviews = pd.read_csv("C:/学习/python/hello/fandango_score_comparison.csv")
 cols = ["FILM", "RT_user_norm", "Metacritic_user_nom", "IMDB_norm", "Fandango_Ratingvalue", "Fandango_Stars"]
 norm_reviews = reviews[cols]
 
 num_cols = ["RT_user_norm", "Metacritic_user_nom", "IMDB_norm", "Fandango_Ratingvalue", "Fandango_Stars"]
 bar_height = norm_reviews.loc[, num_cols].values  #第一部电影的评价，注意利用loc索引某一行的用法，可以添加第二维
 
 bar_position =  + np.arange()  #arange返回的是ndarray类型，range返回的是list类，使用arange需要用numpy
 plt.figure(figsize=(, ))
 ax = plt.subplot(, , )
 ax.barh(bar_position, bar_height, 0.3)  #bar_position是条形图的x坐标(中点坐标),bar_height是高，.3是宽
 #设置x坐标刻度
 tick_positions = range(, )
 ax.set_yticks(tick_positions)
 ax.set_yticklabels(num_cols)
 
 #设置x轴和y轴名称
 ax.set_ylabel("Rating Source")
 ax.set_xlabel("Average Rating")
 ax.set_title("Whatever")
 plt.show()

运行结果如下

6.画散点图

 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 
 reviews = pd.read_csv("C:/学习/python/hello/fandango_score_comparison.csv")
 plt.figure(figsize=(,))
 ax = plt.subplot(, , )
 ax.scatter(reviews["RT_norm"], reviews["Metacritic_user_nom"])
 plt.show()

运行结果如下

7.设a是Series类型，b = a.value_counts()可以得到a的一个频数统计，b是Series结构，b的index是a的值，b的value是该值出现的频数。

如下代码所示

 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 
 reviews = pd.read_csv("C:/学习/python/hello/fandango_score_comparison.csv")
 cols = ["FILM", "RT_user_norm", "Metacritic_user_nom", "Fandango_Ratingvalue"]
 norm_reviews = reviews[cols]
 #fandango_distribution是Series结构，index是原来的列的值，value是该值出现的频率
 fandango_distribution = norm_reviews["Fandango_Ratingvalue"].value_counts()
 print(fandango_distribution.head())
 print(type(fandango_distribution))
 print(fandango_distribution.index)

运行结果如下

8.我们来画直方图

 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 
 reviews = pd.read_csv("C:/学习/python/hello/fandango_score_comparison.csv")
 cols = ["FILM", "RT_user_norm", "Metacritic_user_nom","IMDB_norm", "Fandango_Ratingvalue"]
 norm_reviews = reviews[cols]
 plt.figure(figsize=(, ))
 ax = plt.subplot(, , )
 ax.hist(norm_reviews["RT_user_norm"], bins=) #参数bins表示直方图的x轴分成多少区间
 
 plt.show()

运行结果如下