分组与聚合

Dataframe.groupby(key，as_index=False)

• key:分组的列数据，可以多个

抛开聚合谈分组是没有意义的，只会返回对象

案例

col=pd.DataFrame({'color':['white','red','green','red','green'],'object':['pen','pencil','pen','pencil','ashtray']})

#分组，求平均值
col.groupby(['color'], as_index=False/True)['price'].mean()      #dataframe
col['price'].groupby(col['color']).mean()   #series

Starbucks案例

starbucks=pd.read_csv('./data.csv') #读取数据

starbucks.head() #查看前五行数据

count=starbucks.groupby(['country']).count()  #starbucks.groupby(['country'])不加count（）只会返回对象

#画图
count['brand'].plot(kind='bar',figsize=(20,8))


#也可多个分组
count=starbucks.groupby(['country'，‘state’]).count()