import numpy as npimport pandas as pdimport matplotlib.pyplot as plt%matplotlib inline
bike = pd.read_csv(open(r'D:\python数据分析\数据\bike.csv'))bike.head()

bike.isnull().sum()

bike.dtypes

bike['datetime'] = pd.to_datetime(bike['datetime'])bike.dtypes

bike = bike.set_index('datetime') #将datetime字段设置为DataFrame的索引，成为时间序列数据bike.head()

bike.index #索引

bike.tail()

y_bike = bike.groupby(lambda x: x.year).mean() # 降采样年份数据y_bike['count']

y_bike['count'].plot(kind='bar') # 绘制柱状图

m_bike = bike.resample('M', kind='period').mean() # 重采样到月份，类型为时期类型
m_bike.head()

fig, axes = plt.subplots(2, 1)    #两行一列m_bike['2011']['count'].plot(ax=axes[0],sharex=True)  #贡献X轴m_bike['2012']['count'].plot(ax=axes[1])

bike['day'] = bike.index.daybike['hour'] = bike.index.hour # 单独存储日和时的数据
bike.head()

d_bike = bike.groupby('day')['count'].mean() #对day字段分组统计
d_bike

d_bike.plot() # 自行车每日租赁数分布

h_bike = bike.groupby('hour')['count'].mean() #对hour字段分组统计
h_bike

h_bike.plot() # 自行车每小时租赁数分布

work_bike = bike.groupby('workingday')['count'].mean()
work_bike  #对workingday字段分组统计

work_bike.plot(kind='bar')

weather_bike = bike.groupby('weather')['count'].mean()
weather_bike #对weather字段分组统计

weather_bike.plot(kind='bar')