借网站日记分析~普及一下Pandas基础

%%time

import numpy as np
import pandas as pd

Wall time: 520 ms

%%time

# 生成一个2017年的所有时间点（分钟为单位）
datetime_index = pd.date_range("2017-01-01","2018-01-01",closed="left",freq="min")

Wall time: 4 ms

%%time

# 简单查看一下
datetime_index # 如果你不想取到2018.01.01 00:00:00就设置~闭区间closed="left"

Wall time: 0 ns





DatetimeIndex(['2017-01-01 00:00:00', '2017-01-01 00:01:00',
               '2017-01-01 00:02:00', '2017-01-01 00:03:00',
               '2017-01-01 00:04:00', '2017-01-01 00:05:00',
               '2017-01-01 00:06:00', '2017-01-01 00:07:00',
               '2017-01-01 00:08:00', '2017-01-01 00:09:00',
               ...
               '2017-12-31 23:50:00', '2017-12-31 23:51:00',
               '2017-12-31 23:52:00', '2017-12-31 23:53:00',
               '2017-12-31 23:54:00', '2017-12-31 23:55:00',
               '2017-12-31 23:56:00', '2017-12-31 23:57:00',
               '2017-12-31 23:58:00', '2017-12-31 23:59:00'],
              dtype='datetime64[ns]', length=525600, freq='T')

%%time

# 查看多少元素
datetime_index.size # 用len()也一样

Wall time: 0 ns





525600

%%time

# 创建一个DataFrame，以时间轴为index
log_df = pd.DataFrame(index=datetime_index)

Wall time: 999 µs

%%time

# 随机生成一波访问人数
log_df["访问人数"] = np.random.randint(10,50,size=datetime_index.size)

# 随机生成一波注册人数
log_df["注册人数"] = np.random.randint(3,size=datetime_index.size)

Wall time: 31.2 ms

# 查看前10条
log_df.head(10)

# 查看后10条
log_df.tail(10)

# 查看有多少数据
log_df.size

1051200

# 矩阵形状
log_df.shape # 525601行，2列

(525600, 2)

%%time

# 简明摘要
log_df.info()

<class 'pandas.core.frame.DataFrame'>
DatetimeIndex: 525600 entries, 2017-01-01 00:00:00 to 2017-12-31 23:59:00
Freq: T
Data columns (total 2 columns):
访问人数    525600 non-null int32
注册人数    525600 non-null int32
dtypes: int32(2)
memory usage: 8.0 MB
Wall time: 12 ms

%%time

# 初略统计一下
log_df.describe()

Wall time: 62.5 ms

%%time

from matplotlib import pyplot as plt

Wall time: 179 ms

%%time

p_s = log_df["访问人数"].resample("D").sum() # 平均可以看mean()
r_s = log_df["注册人数"].resample("D").sum()

day_df = pd.DataFrame()
day_df["访问人数"] = p_s
day_df["注册人数"] = r_s

Wall time: 49 ms

day_df.head(10)

# 快速查看一下2017年1月份的统计
day_df["2017-01"]

day_df["2017-01"].index

DatetimeIndex(['2017-01-01', '2017-01-02', '2017-01-03', '2017-01-04',
               '2017-01-05', '2017-01-06', '2017-01-07', '2017-01-08',
               '2017-01-09', '2017-01-10', '2017-01-11', '2017-01-12',
               '2017-01-13', '2017-01-14', '2017-01-15', '2017-01-16',
               '2017-01-17', '2017-01-18', '2017-01-19', '2017-01-20',
               '2017-01-21', '2017-01-22', '2017-01-23', '2017-01-24',
               '2017-01-25', '2017-01-26', '2017-01-27', '2017-01-28',
               '2017-01-29', '2017-01-30', '2017-01-31'],
              dtype='datetime64[ns]', freq='D')

# 比如绘制一份1月份的访问人数一览表
plt.plot(day_df["2017-01"].index,day_df["2017-01"]["访问人数"])
plt.show()

# 比如绘制一份1月份的注册人数一览表
plt.plot(day_df["2017-01"].index,day_df["2017-01"]["注册人数"])
plt.show()

%%time

# 统计一下以月为单位的人数信息
p_s = log_df["访问人数"].resample("M").sum()
r_s = log_df["注册人数"].resample("M").sum()

month_df = pd.DataFrame()
month_df["visitors"] = p_s
month_df["register"] = r_s

Wall time: 53 ms

month_df.tail(10)

# 具体项你可以指定
plt.plot(month_df.index,month_df["visitors"])
plt.show() # 图中3月是低谷

%%time

# 默认是折线图
month_df.plot()

# 密度图
month_df.plot.density()

# 柱形图（叠图，可以计算百分比）
month_df.plot.bar(stacked=True)

# 水平柱形图
month_df.plot.barh() # stacked=True

# 显示一下
plt.show()

Wall time: 720 ms

%%time

# 季度统计
quarter_df = pd.DataFrame()
quarter_df["visitors"] = log_df["访问人数"].resample("Q").sum()
quarter_df["register"] = log_df["注册人数"].resample("Q").sum()

Wall time: 44 ms

# 饼状图
quarter_df.plot.pie(subplots=True,labels=["Spr","Sum","Aut","Win"])
plt.show()