day 08 淘宝数据分析

select * from sdata.dbo.taobaoda

--根据省市统计 cost的总钱数，收藏宝贝的数量，收藏店铺的数量，加入购物车的数量，点击的次数--
select
priov
,city
,SUM(cost) cost
,SUM(baby)*0.2 baby
,SUM(shop)*0.3 shop
,SUM(car)*0.5 car
,SUM(detail) detail
,SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail) sum_total
,COUNT (1) dianji
,(SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail))*100.0/COUNT (1) cover_tate --成交比列--

,case when SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail) <> 0
then SUM(cost)/(SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail) )
else 0 end cost_per
into pdata.dbo.taobao_car_sum_detail_prov
from sdata.dbo.taobaoda
group by
priov
,city

--根据比例看哪个地方的成交率高成交量高--
select * from pdata.dbo.taobao_car_sum_detail_prov
order by cover_tate desc

--全国的总值，对上面的值汇总--
insert into pdata.dbo.taobao_car_sum_detail_prov
select
'全国' priov
,'全国' city
,SUM(cost) cost
,SUM(baby)*0.2 baby
,SUM(shop)*0.3 shop
,SUM(car)*0.5 car
,SUM(detail) detail
,SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail) sum_total
,COUNT (1) dianji
,(SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail))*100.0/COUNT (1) cover_tate --成交比列--

,case when SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail) <> 0
then SUM(cost)/(SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail) )
else 0 end cost_per
from sdata.dbo.taobaoda


--把成本小于全国的，转换率大于全国的筛选出来--
select a.* from pdata.dbo.taobao_car_sum_detail_prov a
left join(select *from
pdata.dbo.taobao_car_sum_detail_prov
where priov ='全国' ) b
on a.cover_tate > b.cover_tate
and a.cost_per <=b.cost_per
where b.cover_tate is not null --b表里面转化率不能为空--
order by sum_total desc --筛选成交量高的--

--根据时间段来统计--
drop table pdata.dbo.taobao_car_sum_detail_prov_hour
select
priov
,city
,hour
,SUM(cost) cost
,SUM(baby)*0.2 baby
,SUM(shop)*0.3 shop
,SUM(car)*0.5 car
,SUM(detail) detail
,SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail) sum_total
,COUNT (1) dianji
,(SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail))*100.0/COUNT (1) cover_tate --成交比列--

,case when SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail) <> 0
then SUM(cost)/(SUM(baby)*0.2 +SUM(shop)*0.3 +SUM(car)*0.5+SUM(detail) )
else 0 end cost_per
into pdata.dbo.taobao_car_sum_detail_prov_hour
from sdata.dbo.taobaoda
group by
priov
,city
,hour

select * from pdata.dbo.taobao_car_sum_detail_prov_hour

--根据哪个时间成交量高分组排序--
select *
into pdata.dbo.orde_by
from
(
select
*
,ROW_NUMBER()over(partition by priov,city order by cover_tate desc) as rows
from pdata.dbo.taobao_car_sum_detail_prov_hour) a
where rows <=4 and cover_tate>10 and cost_per <40

--分析全国在哪个时候成交量最大--
select hour
,count(1) chengjiaoliang
from pdata.dbo.orde_by
group by hour
order by chengjiaoliang desc