Python analyzes the user behavior data of Douyin to see what kind of video will explode!

Preface

The text and pictures in this article are from the Internet and are for learning and communication purposes only. They do not have any commercial use. If you have any questions, please contact us for processing.

Python analysis of Douyin user behavior data video explanation address

https://www.bilibili.com/video/BV1yp4y1q7ZC/Copy 
code

With the rise of short video apps, in the short video social market, Douyin short videos are extremely popular. We hope that through this analysis, we can give users some suggestions for posting videos.

data analysis

1 platform

Daily broadcasts, daily users, daily authors, daily works

 

 

The trend of daily broadcast volume, daily user volume, daily author volume, and daily production volume over time is basically the same: steady growth; during the period from 2019-10-20 to 2019-10-29, all indicators have experienced huge growth first. After that, it tends to be stable, and then falls back to the normal level. It is guessed that at this point in time the platform has carried out promotion activities, so that it has attracted a large number of users.

Author: The number of works, works won praise rate

 

 

The number of author’s works is directly proportional to the broadcast rate

The number of author’s works and the like rate do not have much relationship

Author: Play the amount of contribution

 

 

3,500, 18% or so authors contributed 80% of the broadcast volume of the platform, obeying the 28th rule.

2 works

Source of works

 

 

It can be seen that a large number of sources and channels of works are 0, accounting for 98.48%.

Selected songs top10

 

 

The top ten background music IDs are: 22, 220, 25, 68, 110, 33, 468, 57, 43, 238 (there is no extra information to check the corresponding song name)

 

 

From 2019-10-21 to 2019-10-29 mentioned above, the playback volume of all song works has increased, among which songs with IDs 22, 220, 68, 25 have a skyrocketing trend.

Song and like rate, completion rate

 

Picture uploading

 

There is little difference between the like rate and the completion rate of different background music works, that is, there is little difference between the like and the complete playback result after the playback volume is generated.

 

 

The play volume of different background music works varies greatly, and the play volume of individual songs is outstanding

 

 

Combined with the above picture, most of the songs played on the platform are a small part of popular songs.

 

 

There is little difference in time between the like rate and the completion rate of different songs

The relationship between the duration of different works and the volume of products and playback

 

 

There is a normal proportional relationship between product volume and playback volume of different durations

The product volume (play volume) with a duration of 7-12s accounts for the majority

Play volume above 23s is basically 0

Work duration and completion rate, like rate

 

 

The broadcast rate is generally stable at about 0.4 within 2s-43s, and fluctuates greatly after 43s;

The like rate basically maintains within 0.6 up and down within 2s-43s, and fluctuates greatly after 43s.

Work release time (24H)

 

 

The volume of products in different periods is basically proportional to the volume of playback

10-17 time period, the amount of works and broadcasts on the platform is low (work/study time)

19-0-5 The playback volume of the entire time period is relatively high.

Work release time and completion rate, like rate

 

 

In the 0-5 time period, the completion rate and the like rate of the work are higher

to sum up

Summary of analysis results

platform:

  • Increase event promotion: attract new users and keep old users

  • Increase author incentive projects: encourage authors to publish works

  • Expansion channels: attract new users

Author:

  • Channel: 0

  • Background music: popular songs

  • Work duration: 7-12s, preferably not more than 23s

  • Work release time: 19-0-5 points, of which 0-5 points are better

  • Actively participate in platform activities

  • Recently, many friends consulted about Python learning issues through private messages. To facilitate communication, click on the blue to join the discussion and answer resource base by yourself

 

Guess you like

Origin blog.csdn.net/weixin_43881394/article/details/112605321