1 of 20

PCPP Module 6: Capstone Project in Data Analytics - YouTube Trending Videos

2 of 20

INTRODUCING US ☺

  • Zhi Quan
  • Jeff
  • Denise

​

​

3 of 20

Dataset Details:

  • YOUTUBE TRENDING VIDEOS
  • This dataset includes several months of data on daily trending YouTube videos. Data is provided for the United States, with up to 200 listed trending videos per day.
  • The data also includes a category_id field, which varies between regions.

​

4 of 20

Contents

  • Objective
  • Data quality of the dataset
  • Data Enhancement
  • Youtube Viewer Profile
  • Hypothesis 1
  • Hypothesis 2
  • Hypothesis 3
  • Summary & Recommendation

5 of 20

Objective

  • We, PCPP Group 2, have been assigned to perform exploratory data analysis and data augmentation of youtube trending videos in the United States. Through the analysis, the aim is to provide recommendations and understanding on what drives the numbers behind the platform.

6 of 20

Part One: Exploratory Data Analysis

7 of 20

Data quality of the dataset

  1. Dataset Info
    1. Dataset size: 40,949 records
    2. Total 15 columns with 3 bool, 5 integers, and 7 object type

​

​

2. Data quality

    • Good quality
    • 570 records without description provided
    • Remove columns that we are not using to speed up the data loading process

8 of 20

Data quality of the dataset

9 of 20

Part Two: Data Augmentation

10 of 20

YouTube Viewer Profile

Youtube is the 2nd most popular social media platform with 2.3 Billion monthly active users worldwide. It is also said to be the 2nd most popular search engine after Google which is why it’s key for businesses around the world. People not only get exposed to different kinds of content in video form but with the widely tapped into advertisement by businesses, 90% of people now say they discover new brands and/ or products.

​

​

11 of 20

Hypothesis 1

  • Hypothesis 1: United States of America’s (USA) most popular category on YouTube is Music (category 10) and the least popular category will be the Shows category (category 43).
  • Conclusion: Yes, The greatest number of views are contributed by the Category: Music (category 10) and the least amount is contributed by Shows (category 43).

​

​

​

​

​

12 of 20

Hypothesis 1

Most viewed category: Music

Least viewed category: Shows

13 of 20

Hypothesis 2

  • Hypothesis 2 - Most Popular (views) Category in Youtube will always get Most Rating (likes and dislikes)
  • Conclusion: Yes, the result shows that viewers were generous to express their rating to the content that they like or dislike.

​

14 of 20

Hypothesis 2

15 of 20

Hypothesis 3

  • Hypothesis 3 - The Top 3 Most Popular Youtube Category always received the highest comments (comments count)
  • Conclusion: No, there were significant drops for the 3rd most popular category compared with the 4th, 5th, and 7th most popular categories.

​

16 of 20

Hypothesis 3

17 of 20

Part Three: Conclusion

18 of 20

Summary & Recommendation

  1. There are a total of 43 different categories found in Youtube U.S Dataset, with Music being the Most popular category (41.51%) and surprisingly Shows category is the most unpopular category (0.053%).

​

  • The U.S. population prefer Netflix over YouTube for Shows (Tv, Movie) due to the more advanced platforms in the market. On the other hand, Youtube is the go-to for Music content.

​

  • Youtube viewers in U.S. are generally willing to provide ratings for the videos - like or dislike. This influence the popularity of the channels in the category.

​

  • In terms of comment counts, the trend is not following the popularity of the category. We believe this is mainly due to the content brought by youtuber in the video, if it appeals to the audience and create a topic to discuss about.

​

  • [Recommendation] Using NLP to analyse Youtube tags to understand what kind of content is related to and identified with, in relations to the type of music that is popular within the U.S. population.

​

  • [Recommendation] Businesses can tap into categories that are more popular for advertisements purposes and with tags, businesses are able to make use of that to know what relates to their viewerships better.

​

​

​

19 of 20

References

​

​

​

​

​

20 of 20

Thank You!☺

  • We have come to the end of our presentation, thank you for your kind attention!

​

​