python数据分析：在jupyter notebook上使用python&SQL做数据分析

类似于在jupyter上使用R语言，同样可以使用SQL语句：
详细见github项目：https://github.com/catherinedevlin/ipython-sql

安装ipython-sql

pip install ipython-sql

载入

%load_ext sql

连接数据库 同 SQLAlchemy

• postgresql://will:longliveliz@localhost/shakes
• mysql+pymysql://scott:tiger@localhost/foo
• oracle://scott:[email protected]:1521/sidname
• sqlite://
• sqlite:///foo.db
• mssql+pyodbc://username:password@host/databasedriver=SQL+Server+Native+Client+11.0

我是使用的是mysql，本地链接，用户名ffzs，密码666666，test数据库：

%sql mysql+pymysql://ffzs:666666@localhost/test

简单使用

%matplotlib inline
import matplotlib.pyplot as plt
plt.style.use('bmh')

1.显示表

%%sql
show tables;

2.选取steam_users表的前5行

df = %sql select * from steam_users limit 5
df.DataFrame()

3.计算表中包含多少游戏数和玩家数

%%sql
select count(distinct Game) gameCount, count(distinct UserID) userCount
from steam_users

4.筛选出拥有用户前十的游戏

%%sql
data <<
select Game , count(1) as count from steam_users 
where Action='play'
group by Game 
order by count desc
limit 10

data.DataFrame()[::-1].plot.barh("Game","count")

5.筛选出被玩总时长前十的游戏

%%sql
playHour <<
select Game,sum(Hours) as playHour 
from steam_users 
where Action="play"
group by Game 
order by playHour desc
limit 10

playHour.DataFrame()[::-1].plot.barh('Game', 'playHour')

6.筛选出被玩平均时长前十的游戏

%%sql
avgHour <<
select Game, avg(Hours) as avgHour 
from steam_users 
where Action='play'
group by Game
order by avgHour desc
limit 10

avgHour.DataFrame()[::-1].plot.barh('Game','avgHour')

7.平均时长前十的游戏的游戏人数

%%sql
select Game, avg(Hours) as avgHour, count(1) as count
from steam_users
where Action='play'
group by Game
order by avgHour desc
limit 10

联系join on：

%%sql
select a.Game, avgHour, count from
(select Game, avg(Hours) as avgHour 
from steam_users 
where Action='play'
group by Game
order by avgHour desc
limit 10) a
left join
(select Game ,count(1) as count 
from steam_users 
where Action='play'
group by Game) b
on a.Game=b.Game
order by avgHour desc

可见平均时长长的游戏大多是小众游戏

8.玩家人数大于500人的游戏的个数（having使用）

%%sql
select count(1) as count
from 
(select Game, count(1) as count
from steam_users 
where Action='play' 
group by Game
having count > 500) a

9.拥有游戏数量前十用户

%%sql
games <<
select UserID, count(1) count 
from steam_users
where Action='play'
group by UserID
order by count desc
limit 10

games.DataFrame()[::-1].plot.barh('UserID','count')

10.游戏总时长最多5个用户和最少5个用户（union使用）

%%sql
(select UserID, sum(Hours) as allHour
from steam_users
where Action='play'
group by UserID
order by allHour desc
limit 5)
union
(select UserID, sum(Hours) as allHour
from steam_users
where Action='play'
group by UserID
order by allHour
limit 5)