数据科学第2周 | 基于Python实现数据的描述性统计

配置环境

$ pip3 install jupyterlab  # 安装jupyter
$ jupyter notebook  # 运行jupter

$ pip3 install pandas  # 安装pandas
$ pip3 install matplotlib  # 安装matplotlib

加载数据

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

df = pd.read_csv('salaries_10000.csv')
df.head()

	emp_no	salary	from_date	to_date
0	10001	60117	1986-06-26	1987-06-26
1	10001	62102	1987-06-26	1988-06-25
2	10001	66074	1988-06-25	1989-06-25
3	10001	66596	1989-06-25	1990-06-25
4	10001	66961	1990-06-25	1991-06-25

df.describe()

	emp_no	salary
count	10000.000000	10000.000000
mean	10526.233800	64287.636300
std	304.535669	16940.369764
min	10001.000000	39265.000000
25%	10262.750000	50853.750000
50%	10526.000000	61585.000000
75%	10791.000000	75238.000000
max	11053.000000	136004.000000

数据的集中趋势

# 算术平均数
df['salary'].mean()

64287.6363

# 中位数
df['salary'].median()

61585.0

# 众数 
df['salary'].mode()

0    40000
dtype: int64

# 最小值与最大值
print("最小值", df['salary'].min()) 
print("最大值", df['salary'].max()) 

最小值 39265
最大值 136004

# 分位数
print ("第一分位数: ", df['salary'].quantile(q=0.25))
print ("第二分位数: ", df['salary'].quantile(q=0.50))
print ("第三分位数: ", df['salary'].quantile(q=0.75))

第一分位数:  50853.75
第二分位数:  61585.0
第三分位数:  75238.0

数据的离中趋势

# 方差
df['salary'].var()

286976127.73639596

# 标准差
df['salary'].std()

16940.36976386277

# 四分位差
df['salary'].quantile(q=0.75)-df['salary'].quantile(q=0.25)

24384.25

# 离散系数：标准差 / 均值
df['salary'].std() / df['salary'].mean()

0.2635089846017993

plt.hist(df['salary'],100)

(array([390., 105., 150., 166., 231., 180., 177., 220., 231., 210., 222.,
        223., 233., 228., 228., 247., 210., 245., 234., 206., 192., 227.,
        235., 205., 216., 190., 193., 200., 189., 167., 171., 174., 163.,
        165., 160., 143., 153., 119., 136., 119., 123., 139., 106., 118.,
        106.,  89., 111.,  92.,  98.,  83.,  80.,  62.,  74.,  79.,  69.,
         74.,  58.,  66.,  52.,  55.,  52.,  37.,  36.,  31.,  30.,  28.,
         21.,  22.,  17.,  22.,   5.,  12.,  12.,  12.,   4.,   8.,   6.,
          6.,   4.,   3.,   7.,   6.,   7.,   2.,   5.,   2.,   2.,   2.,
          2.,   5.,   1.,   1.,   0.,   0.,   0.,   1.,   0.,   0.,   0.,
          2.]),
 array([ 39265.  ,  40232.39,  41199.78,  42167.17,  43134.56,  44101.95,
         45069.34,  46036.73,  47004.12,  47971.51,  48938.9 ,  49906.29,
         50873.68,  51841.07,  52808.46,  53775.85,  54743.24,  55710.63,
         56678.02,  57645.41,  58612.8 ,  59580.19,  60547.58,  61514.97,
         62482.36,  63449.75,  64417.14,  65384.53,  66351.92,  67319.31,
         68286.7 ,  69254.09,  70221.48,  71188.87,  72156.26,  73123.65,
         74091.04,  75058.43,  76025.82,  76993.21,  77960.6 ,  78927.99,
         79895.38,  80862.77,  81830.16,  82797.55,  83764.94,  84732.33,
         85699.72,  86667.11,  87634.5 ,  88601.89,  89569.28,  90536.67,
         91504.06,  92471.45,  93438.84,  94406.23,  95373.62,  96341.01,
         97308.4 ,  98275.79,  99243.18, 100210.57, 101177.96, 102145.35,
        103112.74, 104080.13, 105047.52, 106014.91, 106982.3 , 107949.69,
        108917.08, 109884.47, 110851.86, 111819.25, 112786.64, 113754.03,
        114721.42, 115688.81, 116656.2 , 117623.59, 118590.98, 119558.37,
        120525.76, 121493.15, 122460.54, 123427.93, 124395.32, 125362.71,
        126330.1 , 127297.49, 128264.88, 129232.27, 130199.66, 131167.05,
        132134.44, 133101.83, 134069.22, 135036.61, 136004.  ]),
 <a list of 100 Patch objects>)

print ("偏态系数: ", df['salary'].skew())
print ("峰态系数: ", df['salary'].kurt())

偏态系数:  0.6784771437175181
峰态系数:  -0.03592105965629688