Barbie q! How to analyze the new game app?

【 subject 】

In order to monitor the newly listed Games APP Popularity of , Analyze the total number of users through data 、 Average age of users and active users （ Visit for two consecutive days ） The total number and average age of . The following table shows the details of user login information table .

Field explanation ：

Login date ： Record user login to the game APP Time for

user ID： The unique identification of the user

User age ： The user is in APP Age of registration

The following results need to be analyzed ：

1. Count the total number of users 、 Average age of users

2. Count active users （ Visit for two consecutive days ） The total number and average age of

【 Their thinking 】

1. Count the total number of users 、 Average age of users

From table 1, you can see that the same user has multiple login records on the same day , If you perform aggregate queries directly, it will cause duplicate data calculation , So you should first follow the user ID The de emphasis of duplicate data is to analyze the total number and average age of users .

SQL Writing a ：

select distinct  user ID, User age 
from  User login information table ;

SQL Write two ：

select   user ID, User age 
from  User login information table   
group by  user ID;

Query results ：

The query results are used as temporary tables a, Next, use to calculate the number of users （count function ） And ask the average age of users （avg function ）

SQL It is written as follows ：

select  count( user ID) as  Total users  ,avg( User age ) as  Average age of users 
from(
select   user ID, User age 
from  User login information table 
group by  user ID
) a;

Query results ：

2. Count active users （ Visit for two consecutive days ） The total number and average age of

Active users ： It is positioned as having access for two consecutive days , The vernacular is interpreted as logging in today , Number of users logged in tomorrow . That is, the time interval =1.

If a table involves time intervals , You need to use self connection , That is to join two identical tables . Will table a Understood as the user's login time , Will table b Understood as the user's login time again , By calculating the time difference between login time and login time again , The time difference is equal to 1 That means you are also logged in today , Log in tomorrow （ Log in for two consecutive days ）

select
        a. Login date  as  The login time ,
        a. user ID,
        a. User age ,
        b. Login date  as  Login time again 
from
         User login information table  as a
left join   User login information table  as b on a. user ID = b. user ID;

Query results ：

The temporary table after connection is marked as table c, So how to start from the table c Find out the time interval （ Landing time tomorrow - Today's landing time ）=1 What about the data? ？

（1） This involves calculating the difference between two dates ,《 monkey Learn from scratch sql》 The corresponding single function mentioned in is timestampdiff. The following figure shows the usage of this function .

SQL Write the following ：

select *,timestampdiff(day,c. The login time ,c. Logout time ) as  The time interval 
From c
group by c. user ID,c. The login time ;

Query results ：

The query results are used as temporary tables d, use case when Statement selects the time interval =1 The data of , Calculate the total number of active users （count） And the average age of active users （avg）

SQL Write the following ：

select count(distinct case when  The time interval =1 then  user ID else null end ) as  Total active users ,
avg( case when  The time interval =1 then  User age  else null end ) as   Average age of active users 
from d;

Put the temporary table d Of SQL Plug in , Final SQL The code is as follows ：

select
count(distinct case when  The time interval =1 then  user ID else null end ) as  Total active users ,
avg( case when  The time interval =1 then  User age  else null end ) as   Average age of active users 
from
(
select *,timestampdiff(day,c. The login time ,c. Login time again ) as  The time interval 
from
(
select
       a. Login date  as  The login time ,
       a. user ID,
       a. User age ,
       b. Login date  as  Login time again 
from User login information table  as a
left join   User login information table  as b on a. user ID = b. user ID
) c
group by c. user ID,c. The login time 
) d;

Query results ：

【 The test point of this question 】

1. Understanding of common indicators , For example, the number of retained users 、 Retention .

2. Flexible use case To statistics when Function and group by Perform custom column statistics .

3. Encountered only one table , But the problem of counting time intervals , We need to think of using self connection to find the time interval , Similar to finding continuity appear N The content of this time 、 sound of dripping water 2020 True job title 、 Chain home interview questions ： How to analyze retention .

▼  Click on 「 Read the original 」

▼ Unlock more data analysis courses