在閱讀前，讀一下Python-深入淺出數據分析-總結會更好點，以后遇到問題比如代碼運行不了，再讀讀也行，>-_-<
數字的分布
首先書中數落的一下Excel的直方圖功能，其實呢，Excel2016已經改進了很多了
Python中的直方圖和箱形圖
用不了幾行代碼，是不是比Excel簡單美觀點
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
df = pd.read_csv('./hfda_ch09_employees.csv', skiprows=1, names =['staff_num', 'received', 'negotiated', 'gender', 'year'])
fig = plt.figure(figsize=(12, 6))
ax1 = fig.add_subplot(1, 3, 1)
l = ax1.boxplot(df['received'].values)
ax2 = fig.add_subplot(1, 2, 2)
l = ax2.hist(df['received'], bins=50)
各種情況的對比
分析不同維度的數字分布
fig, ((ax1, ax2), (ax3, ax4), (ax5, ax6)) = plt.subplots(nrows=3, ncols=2, figsize=(16,16))
ax1.hist(df['received'][df['year']==2007], bins=50)
ax1.set_title('year=2007')
ax2.hist(df['received'][df['year']==2008], bins=50)
ax2.set_title('year=2008')
ax3.hist(df['received'][df['gender']=='M'], bins=50)
ax3.set_title('gender=M')
ax4.hist(df['received'][df['gender']=='F'], bins=50)
ax4.set_title('gender=F')
ax5.hist(df['received'][df['negotiated']], bins=50)
ax5.set_title('negotiated=TRUE')
ax6.hist(df['received'][~df['negotiated']], bins=50)
ax6.set_title('negotiated=FALSE')
