商业智能（BI）实战项目

期待您的关注

☀大数据学习笔记

1.实现的功能

2.数据库操作步骤

java 复制代码

创建数据库：

create database card;

创建表：

create table card_apply
(
cid bigint primary key auto_increment
,apply_uid bigint
,apply_entrance varchar(90)
,apply_dt varchar(90)
,access_state varchar(90)
,issuccess varchar(90)
,source varchar(90)
,isact varchar(10)
,isfirst varchar(10)
);

环比：这个月减去上个月的和上个月的比

同比：今年本月减去去年同期和去年同月的比

java 复制代码

插入数据：

insert into card_apply values
(null,1,'广告banner','2024-06-04','点击','不成功','直销','否','否')
,(null,2,'首页con','2024-06-04','点击','不成功','直销','否','否')
,(null,3,'交叉销售广告位','2024-06-04','点击','不成功','直销','否','否')
,(null,4,'信用卡专区','2024-05-02','点击','不成功','直销','否','否')
,(null,5,'交叉销售广告位','2024-05-03','点击','不成功','直销','否','否')
,(null,6,'广告banner','2024-05-03','点击','不成功','直销','否','否')
,(null,7,'广告banner','2023-06-03','点击','成功','直销','是','是')
,(null,8,'交叉销售广告位','2023-06-03','点击','成功','直销','是','是');


select * from card_apply;

3.计算本月信用卡申请人数及同比环比

打开Tableau 连接MySQL 双击新自定义SQL 写入SQL语句

java 复制代码

select
a.user_cnt curr_month_user_cnt,
(a.user_cnt-b.user_cnt)/b.user_cnt upper_month_user_cnt,
(a.user_cnt-c.user_cnt)/c.user_cnt last_year_user_cnt
from
(
select count(distinct(apply_uid)) user_cnt from card_apply where date_format(apply_dt,'%Y-%m')=date_format(current_date(),'%Y-%m')
)a
,(
select count(distinct(apply_uid)) user_cnt from card_apply where date_format(apply_dt,'%Y-%m')=date_format(date_sub(current_date(), interval 1 month),'%Y-%m')
)b
,(
select count(distinct(apply_uid)) user_cnt from card_apply where date_format(apply_dt,'%Y-%m')=date_format(date_sub(current_date(), interval 12 month),'%Y-%m')
)c

MySQL日期函数

然后跳转到工作表

4.计算本月信用卡核卡成功人数及同比环比

java 复制代码

为了方便计算 我们再插入几条数据：

insert into card_apply values
(null,9,'交叉销售广告位','2024-05-03','点击','成功','直销','是','是'),
(null,10,'信用卡专区','2024-06-03','点击','成功','直销','是','是'),
(null,11,'首页con','2024-05-08','点击','成功','直销','是','是');

打开Tableau 连接MySQL 双击新自定义SQL 写入SQL语句

java 复制代码

SELECT
a.success_uid current_success_uid
,(a.success_uid-b.success_uid)/b.success_uid last_month_success_uid
,(a.success_uid-c.success_uid)/c.success_uid last_year_success_uid
FROM
(
select count(distinct(apply_uid)) success_uid from card_apply where issuccess = '成功' and date_format(apply_dt,'%Y-%M') = date_format(current_date(),'%Y-%M')
)a
,(
select count(distinct(apply_uid)) success_uid from card_apply where issuccess = '成功' and date_format(apply_dt,'%Y-%M') = date_format(date_sub(current_date(),interval 1 month),'%Y-%M')
)b
,(
select count(distinct(apply_uid)) success_uid from card_apply where issuccess = '成功' and date_format(apply_dt,'%Y-%M') = date_format(date_sub(current_date(),interval 12 month),'%Y-%M')
)c

最终结果

5.计算本月信用卡激活人数及同比环比

java 复制代码

为了方便计算 我们再插入几条数据：

insert into card_apply values
(null,12,'交叉销售广告位','2024-06-23','点击','成功','直销','是','是'),
(null,13,'信用卡专区','2024-06-18','点击','成功','直销','是','是'),
(null,14,'首页con','2024-05-08','点击','不成功','直销','是','是');

打开Tableau 连接MySQL 双击新自定义SQL 写入SQL语句

java 复制代码

SELECT
a.act_uid current_act_uid
,(a.act_uid-b.act_uid)/b.act_uid last_month_act_uid
,(a.act_uid-c.act_uid)/c.act_uid last_year_act_uid
FROM
(
select count(distinct(apply_uid)) act_uid from card_apply where isact = '是' and date_format(apply_dt,'%Y-%M') = date_format(current_date(),'%Y-%M')
)a
,(
select count(distinct(apply_uid)) act_uid from card_apply where isact = '是' and date_format(apply_dt,'%Y-%M') = date_format(date_sub(current_date(),interval 1 month),'%Y-%M')
)b
,(
select count(distinct(apply_uid)) act_uid from card_apply where isact = '是' and date_format(apply_dt,'%Y-%M') = date_format(date_sub(current_date(),interval 12 month),'%Y-%M')
)c

最终结果

6.计算本月信用卡首刷人数及同比环比

java 复制代码

为了方便计算 我们再插入几条数据：


insert into card_apply values
(null,15,'交叉销售广告位','2024-06-26','点击','成功','直销','是','否'),
(null,16,'信用卡专区','2024-06-11','点击','成功','直销','是','是'),
(null,17,'首页con','2023-05-08','点击','不成功','直销','是','是');

打开Tableau 连接MySQL 双击新自定义SQL 写入SQL语句

java 复制代码

SELECT
a.first_uid current_first_uid
,(a.first_uid-b.first_uid)/b.first_uid last_month_first_uid
,(a.first_uid-c.first_uid)/c.first_uid last_year_first_uid
FROM
(
select count(distinct(apply_uid)) first_uid from card_apply where isfirst = '是' and date_format(apply_dt,'%Y-%M') = date_format(current_date(),'%Y-%M')
)a
,(
select count(distinct(apply_uid)) first_uid from card_apply where isfirst = '是' and date_format(apply_dt,'%Y-%M') = date_format(date_sub(current_date(),interval 1 month),'%Y-%M')
)b
,(
select count(distinct(apply_uid)) first_uid from card_apply where isfirst = '是' and date_format(apply_dt,'%Y-%M') = date_format(date_sub(current_date(),interval 12 month),'%Y-%M')
)c

最终结果

7.求本月信用卡来源

java 复制代码

为了方便计算 我们再插入几条数据：

insert into card_apply values
(null,15,'交叉销售广告位','2024-06-29','点击','成功','网络','是','是'),
(null,16,'信用卡专区','2024-06-18','点击','成功','网络','是','是'),
(null,17,'首页con','2024-06-06','点击','不成功','网络','是','是'),
(null,18,'交叉销售广告位','2024-06-21','点击','成功','分行','是','是'),
(null,19,'信用卡专区','2024-06-30','点击','成功','分行','是','是'),
(null,20,'首页con','2024-06-09','点击','不成功','分行','是','是'),
(null,21,'信用卡专区','2024-06-30','点击','成功','网络外部平台','是','是'),
(null,22,'首页con','2024-06-09','点击','不成功','网络外部平台','是','是'),
(null,23,'信用卡专区','2024-06-30','点击','成功','网络外部平台','是','是'),
(null,24,'首页con','2024-06-09','点击','不成功','网络外部平台','是','是');

java 复制代码

SELECT
count(*)
,source
FROM card_apply where date_format(apply_dt,'%Y-%M')=date_format(current_date(),'%Y-%M')
group by source;

8.求本月信用卡核卡成功用户来源

java 复制代码

SELECT
COUNT(*),source
FROM card_apply
WHERE DATE_FORMAT(apply_dt,'%Y-%M')=DATE_FORMAT(CURRENT_DATE(),'%Y-%M') and issuccess = '成功'
GROUP BY source

9.求本月APP信用卡申请入口分布

java 复制代码

SELECT
COUNT(*),apply_entrance
FROM card_apply
WHERE DATE_FORMAT(apply_dt,'%Y-%M')=DATE_FORMAT(CURRENT_DATE(),'%Y-%M')
GROUP BY apply_entrance

10.求本月APP信用卡申请和提交用户数

java 复制代码

为了方便计算 我们再插入几条数据：


insert into card_apply values
(null,25,'交叉销售广告位','2024-06-26','提交','成功','直销','是','否'),
(null,26,'信用卡专区','2024-06-11','提交','成功','直销','是','是'),
(null,27,'首页con','2023-05-08','提交','不成功','直销','是','是'),
(null,28,'交叉销售广告位','2024-06-25','提交','成功','直销','是','否'),
(null,29,'信用卡专区','2024-06-25','提交','成功','直销','是','是'),
(null,30,'首页con','2023-06-09','提交','不成功','直销','是','是'),
(null,31,'交叉销售广告位','2024-06-25','提交','成功','直销','是','否'),
(null,32,'信用卡专区','2024-06-24','提交','成功','直销','是','是'),
(null,33,'首页con','2023-06-09','提交','不成功','直销','是','是'),
(null,34,'交叉销售广告位','2024-06-18','提交','成功','直销','是','否'),
(null,35,'信用卡专区','2024-06-19','提交','成功','直销','是','是'),
(null,36,'首页con','2023-06-09','提交','不成功','直销','是','是'),
(null,37,'首页con','2023-06-21','提交','不成功','直销','是','是'),
(null,38,'首页con','2023-06-22','提交','不成功','直销','是','是'),
(null,39,'首页con','2023-06-23','提交','不成功','直销','是','是'),
(null,40,'首页con','2023-06-24','提交','不成功','直销','是','是'),
(null,41,'首页con','2023-06-20','提交','不成功','直销','是','是')

java 复制代码

SELECT
access_state
,apply_dt
,count(distinct(apply_uid)) user_cnt
FROM card_apply
WHERE DATE_FORMAT(apply_dt,'%Y-%M')=DATE_FORMAT(CURRENT_DATE(),'%Y-%M')
and (access_state = '点击' or access_state = '提交')
group by apply_dt,access_state

由于数据量较小所以曲线不是那么圆顺如果我们想要让它变得更丝滑那么我们可以添加大量数据

11.汽车服务用户分析的设计与实现

java 复制代码

create database app;

use app;


create table app_log
(
log_id bigint primary key auto_increment
,access_dt varchar(90)
,user_id bigint
,browse_page varchar(10)
,app_status varchar(90)
,operator varchar(30)
,net_type varchar(20)
);

java 复制代码

插入数据：

insert into app_log values 
(null,'2024-06-01',1,'A','异常','移动','4G'),
(null,'2024-06-02',2,'B','异常','移动','4G'),
(null,'2024-06-01',2,'C','异常','移动','4G'),
(null,'2024-06-02',3,'D','异常','移动','4G'),
(null,'2024-06-02',4,'D','异常','移动','4G')

1）求每日新增用户数

那么该如何求新增用户数呢（使用连接查询）

java 复制代码

select 
'2024-06-02' dt
,count(distinct(a.user_id)) num
from 
(
select user_id from app_log where date_format(access_dt,'%Y-%m-%d') = '2024-06-02'
)a 
left join 
(
select user_id from app_log where date_format(access_dt,'%Y-%m-%d') < '2024-06-02'
)b
on a.user_id = b.user_id WHERE b.user_id is null

2）求新增用户7日留存率

日期加减：date_sub(current_date(),interval 8 day)

java 复制代码

select 
c.dt,
c.user_cnt/e.user_cnt remain_ration
from
(
select 
date_sub('2024-06-02',interval 7 day) dt,
count(distinct(a.user_id)) user_cnt
from 
(
select user_id from app_log where date_format(access_dt,'%Y-%m-%d') = '2024-06-02'
)a 
inner join 
(
select user_id from app_log where date_format(access_dt,'%Y-%m-%d') = date_sub('2024-06-02',interval 8 day)
)b
on a.user_id = b.user_id
)c 
,
(
select count(distinct(user_id)) user_cnt
from 
(
select user_id from app_log where date_format(access_dt,'%Y-%m-%d') = date_sub('2024-06-02',interval 8 day)
)b
)e

3）求每周新增用户数

日期转周：week(current_date())

java 复制代码

select
count(distinct(a.user_id)) user_cnt
from 
(
select 
distinct(user_id) user_id
from app_log where week(access_dt) = 23
)a 
left join 
(
select 
distinct(user_id) user_id
from app_log where week(access_dt) = 22
)b 
on a.user_id = b.user_id where b.user_id is null

4）求新增用户八周留存

java 复制代码

select
d.week,
d.user_cnt/e.user_cnt remain_ration
from 
(
select 
week(date_add(current_date(),interval -9 week)) week,
count(distinct(a.user_id)) user_cnt
from 
(
select user_id from app_log where week(access_dt)=week(current_date())
)a 
inner join 
(
select user_id from app_log where week(access_dt)=week(date_add(current_date(),interval -9 week))
)b on a.user_id = b.user_id
)d,
(
select count(distinct(b.user_id)) user_cnt
from
(
select user_id from app_log where week(access_dt)=week(date_add(current_date(),interval -9 week))
)b
)e

5）求30日活跃用户

java 复制代码

select
count(distinct(user_id)) user_cnt
from app_log 
where access_dt between date_add(current_date(),interval -30 day) and current_date()