机器学习基础——matplotlib.pyplot和seaborn的使用

import matplotlib.pyplot as plt
import numpy as np

第一步 生成数据集

x = np.linspace(-3,3,50)#平均采样,[-3,3]采样50个

x.shape

(50,)

y1 = 2*x + 1

y1.shape

(50,)

y2 = x**2
y2

array([9.00000000e+00, 8.28029988e+00, 7.59058726e+00, 6.93086214e+00,
       6.30112453e+00, 5.70137443e+00, 5.13161183e+00, 4.59183673e+00,
       4.08204915e+00, 3.60224906e+00, 3.15243648e+00, 2.73261141e+00,
       2.34277384e+00, 1.98292378e+00, 1.65306122e+00, 1.35318617e+00,
       1.08329863e+00, 8.43398584e-01, 6.33486047e-01, 4.53561016e-01,
       3.03623490e-01, 1.83673469e-01, 9.37109538e-02, 3.37359434e-02,
       3.74843815e-03, 3.74843815e-03, 3.37359434e-02, 9.37109538e-02,
       1.83673469e-01, 3.03623490e-01, 4.53561016e-01, 6.33486047e-01,
       8.43398584e-01, 1.08329863e+00, 1.35318617e+00, 1.65306122e+00,
       1.98292378e+00, 2.34277384e+00, 2.73261141e+00, 3.15243648e+00,
       3.60224906e+00, 4.08204915e+00, 4.59183673e+00, 5.13161183e+00,
       5.70137443e+00, 6.30112453e+00, 6.93086214e+00, 7.59058726e+00,
       8.28029988e+00, 9.00000000e+00])

plt.figure()
plt.plot(x,y1)

[]

output_7_1.png

plt.figure()
plt.plot(x,y2)

[]

output_8_1.png

plt.plot(x,y2)
plt.show()

output_9_0.png

# 将x,y1,y2画在一起
plt.plot(x,y1)
plt.plot(x,y2)

[]

output_11_1.png

支持中文字体

from pylab import mpl#import matplotlib as mpl
mpl.rcParams['font.sans-serif']=['FangSong']
mpl.rcParams['axes.unicode_minus']=False

# 参数修改
plt.plot(x,y1,'.b')
plt.plot(x,y2,color='r',linewidth=5.0,linestyle=':')#linestyle取值："-"，”-.“，":".该变量是复合变量也可以省略字段直接写”.r“

[]

output_14_1.png

##label标记

plt.plot([1,2,3,4],[2,3,3,3])
plt.ylabel('Some Num')
plt.xlabel('自变量')#默认不支持中文字体

Text(0.5,0,'自变量')

output_16_1.png

散点图

plt.plot([1,2,3,4],[2,3,3,3],'g^')

[]

output_18_1.png

常用的linestyle

ro:红色的圆点

bs:蓝色的方块

g^:绿色的三角

t=np.linspace(-5,5,100)

plt.plot(t,t**2)
plt.plot(t,t**5)

[]

output_21_1.png

plt.plot(t,t**2,'r--',t,t**5,'y-.')#多个函数图，可以合并为一个函数，但是要求（自变量,因变量，style字段）

[,
 ]

output_22_1.png

结构化数据绘制散点图

np.arange(50)

array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
       17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33,
       34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49])

data = {
    'a':np.arange(50),
    'c':np.random.randint(0,50,50),
    'd':np.random.rand(50)
}
data

{'a': array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
        17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33,
        34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49]),
 'c': array([21, 22, 31,  1, 30, 13, 47, 19, 16, 45, 45, 34, 24, 11, 30, 49,  3,
        38, 24, 26,  9, 24, 33, 44, 48, 49,  6, 49,  8, 30, 11, 43, 16, 25,
        29, 34, 14, 21,  4, 20, 13, 46, 11, 25, 20, 39, 41, 34, 47, 36]),
 'd': array([0.03337497, 0.58555231, 0.6983719 , 0.3098672 , 0.0355206 ,
        0.27251523, 0.968375  , 0.7585922 , 0.53316131, 0.2134523 ,
        0.76735142, 0.56798347, 0.98154299, 0.07708504, 0.93535569,
        0.84546409, 0.13395731, 0.24076688, 0.44660032, 0.88671819,
        0.00921326, 0.39650877, 0.44355761, 0.30306934, 0.98691421,
        0.39195663, 0.6424303 , 0.68474638, 0.02455291, 0.90485831,
        0.7171299 , 0.18596694, 0.12510926, 0.57805232, 0.93718472,
        0.21482173, 0.02909599, 0.26395894, 0.39508085, 0.74490499,
        0.17457859, 0.93607408, 0.58727838, 0.76517609, 0.53999965,
        0.5932926 , 0.05968155, 0.70313421, 0.72178338, 0.47063122])}

plt.scatter()绘制散点图

plt.scatter('a','d',data=data)
plt.xlabel('a 数据')
plt.ylabel('d 数据')

Text(0,0.5,'d 数据')

output_27_1.png

plt.scatter('a','c',data=data)

output_28_1.png

data['b'] = np.abs(data['d'])

plt.scatter('a','b',data = data,marker='>',c = 'c')

output_30_1.png

plt.scatter('c','d',data = data,marker='>',c = 'c')

output_31_1.png

data

{'a': array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
        17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33,
        34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49]),
 'c': array([21, 22, 31,  1, 30, 13, 47, 19, 16, 45, 45, 34, 24, 11, 30, 49,  3,
        38, 24, 26,  9, 24, 33, 44, 48, 49,  6, 49,  8, 30, 11, 43, 16, 25,
        29, 34, 14, 21,  4, 20, 13, 46, 11, 25, 20, 39, 41, 34, 47, 36]),
 'd': array([0.03337497, 0.58555231, 0.6983719 , 0.3098672 , 0.0355206 ,
        0.27251523, 0.968375  , 0.7585922 , 0.53316131, 0.2134523 ,
        0.76735142, 0.56798347, 0.98154299, 0.07708504, 0.93535569,
        0.84546409, 0.13395731, 0.24076688, 0.44660032, 0.88671819,
        0.00921326, 0.39650877, 0.44355761, 0.30306934, 0.98691421,
        0.39195663, 0.6424303 , 0.68474638, 0.02455291, 0.90485831,
        0.7171299 , 0.18596694, 0.12510926, 0.57805232, 0.93718472,
        0.21482173, 0.02909599, 0.26395894, 0.39508085, 0.74490499,
        0.17457859, 0.93607408, 0.58727838, 0.76517609, 0.53999965,
        0.5932926 , 0.05968155, 0.70313421, 0.72178338, 0.47063122]),
 'b': array([0.03337497, 0.58555231, 0.6983719 , 0.3098672 , 0.0355206 ,
        0.27251523, 0.968375  , 0.7585922 , 0.53316131, 0.2134523 ,
        0.76735142, 0.56798347, 0.98154299, 0.07708504, 0.93535569,
        0.84546409, 0.13395731, 0.24076688, 0.44660032, 0.88671819,
        0.00921326, 0.39650877, 0.44355761, 0.30306934, 0.98691421,
        0.39195663, 0.6424303 , 0.68474638, 0.02455291, 0.90485831,
        0.7171299 , 0.18596694, 0.12510926, 0.57805232, 0.93718472,
        0.21482173, 0.02909599, 0.26395894, 0.39508085, 0.74490499,
        0.17457859, 0.93607408, 0.58727838, 0.76517609, 0.53999965,
        0.5932926 , 0.05968155, 0.70313421, 0.72178338, 0.47063122])}

柱状图

names = ['A类型','B类型','C类型']
value = [1,10,100]

plt.bar(range(len(names)),value)
plt.xticks(range(len(names)),names)#横坐标

([,
  ,
  ],
 )

output_35_1.png

plt.scatter(names,value)

output_36_1.png

plt.scatter(range(len(names)),value)
plt.xticks(range(len(names)),names)

([,
  ,
  ],
 )

output_37_1.png

plt.scatter(range(len(names)),value)
plt.xticks(range(len(names)),names)
plt.title('离散数据散点图')

Text(0.5,1,'离散数据散点图')

output_38_1.png

子图 SubPlot

1. 讲一个画布进行切分(Figure)

2.将切分后的图分配到固定的位置

3.将图可以设置成固定的大小

plt.figure(1)
plt.subplot(131)#一行三列放在第一的位置
plt.bar(names,value,color='r')
plt.subplot(235)#二行三列放在第五的位置
plt.scatter(names,value,color='y')
plt.subplot(233)#二行三列放在第三的位置
plt.plot(names,value,color='g')
plt.title("离散数据的柱状图，散点图，折线图")

Text(0.5,1,'离散数据的柱状图，散点图，折线图')

output_40_1.png

第 2 部分 Seaborn的绘图练习

道/法/术/器

import seaborn as sns

tips = sns.load_dataset('tips')
tips

# total_bill 和 

2.1 带状图-离散数据和l连续数据的之间的关系

sns.stripplot(data=tips,x='day',y='total_bill',jitter = True)#jitter抖动，默认为TRUE

output_47_1.png

蜂群图-离散数据和连续数据之间的关系-密度排列

sns.swarmplot(x='day',y='total_bill',data=tips)

output_49_1.png

tips.head()

分析每天中 午餐和晚餐的账单分布

2.3 Hue 分组参数

sns.swarmplot(x='day',y='total_bill',data=tips,hue='time')

output_53_1.png

在每天的付账人群中的性别分布

sns.swarmplot(x='day',y='total_bill',data=tips,hue='sex')

output_55_1.png

sns.swarmplot(x='day',y='total_bill',data=tips,hue='size')

output_56_1.png

sns.swarmplot(x='size',y='total_bill',data=tips)

output_57_1.png

# 上图解释了，pizza的不同size的基础价格

# size和相关系数(皮尔逊系数)的关系
tips['size'].corr(tips['total_bill'])

0.5983151309049012

2.4箱型图

sns.boxplot('day','total_bill',data=tips)

output_61_1.png

sns.swarmplot('day','total_bill',data=tips)

output_63_1.png

sns.boxplot("day","total_bill",data=tips,hue='time')

output_64_1.png

2.5 提琴图

如何来表示total_bill的概率分布

sns.violinplot('day','total_bill',data=tips,hue='time')

/anaconda3/lib/python3.7/site-packages/scipy/stats/stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval

output_66_2.png

sns.violinplot('day','total_bill',data=tips,hue='time',split = True)

/anaconda3/lib/python3.7/site-packages/scipy/stats/stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval

output_67_2.png

### 多图合成展示

sns.violinplot('day','total_bill',data=tips)
sns.swarmplot('day','total_bill',data=tips,color='w')

/anaconda3/lib/python3.7/site-packages/scipy/stats/stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval

output_69_2.png

3.0 单一变量估计---离散型变量的统计

每天的交易数量

sns.countplot('day',data=tips)

output_73_1.png

sns.countplot('time',data=tips)

output_74_1.png

sns.countplot('day',data=tips,hue='time')

output_75_1.png

4.0 连续型数据的核密度估计

tips.head()

sns.distplot(tips['total_bill'])

/anaconda3/lib/python3.7/site-packages/scipy/stats/stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval

output_78_2.png

抵消偏度(修改接近正态分布)

sns.distplot(np.log(tips['total_bill']))

/anaconda3/lib/python3.7/site-packages/scipy/stats/stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval

output_80_2.png

sns.distplot(np.sqrt(tips['total_bill']))

/anaconda3/lib/python3.7/site-packages/scipy/stats/stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval

output_81_2.png

抽取前99.5%的数据（去除离散值的方法）

np.percentile(tips['total_bill'],99.5)

48.317099999999996

tips[tips['total_bill']>48.31]

带回归的散点图

sns.lmplot('size','total_bill',data = tips)

/anaconda3/lib/python3.7/site-packages/scipy/stats/stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval

output_86_2.png

联合分布（既有散点的特性，又有线性回归，同事还有概率分布）

sns.jointplot('total_bill','tip',data=tips,kind='reg')

/anaconda3/lib/python3.7/site-packages/scipy/stats/stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval

output_88_2.png