In today's age, where most people have access to a variety of electronic devices, more and more people are consuming content using online video services, among which YouTube is the most popular. According to Google, which owns YouTube, over 400 hours of video are uploaded every minute, 100s of millions of hours are watched each day, and billions of users visit YouTube each month. Thus, YouTube provides something of interest for creators, viewers, and advertisers. YouTube videos are monetized through ad revenues, so the more views a video receives, the more ad revenue that can be generated. From the advertiser's point of view, they want to market their product to as wide an audience as possible, so they will want their ads on the most popular videos. Moreover, currently COVID-19 has caused large numbers of people to stay at home, which will undoubtedly cause a spike in the number of daily YouTube visitors.
In our project, we are analyzing data on daily trending YouTube videos. Given YouTube’s large library of videos, it is important to filter videos to different users. YouTube uses a combination of different factors including measuring users interactions (number of views, shares, comments and likes) to determine if a video is trending. These trending videos are then made visible to users so they can see which videos are popular at a given moment. We are looking at what factors make a video become trending and if these factors are different across various countries and categories. We are also exploring why certain videos generate more views, likes and comments.
Some of the research questions we are looking at are:
- What are common factors among trending videos?
- What differences are there between trending videos in different regions (ex between United States, Great Britain and India)?
- Are there factors that are generalizable across regions?
- Do videos with more likes or more comments get more views?
- Can we categorize the videos based on comments and other engagement metrics?
- What are the most popular categories among trending videos?
- On a given trending day, would there be more dislikes than likes for any video in general?
- What time of the day would be the ideal to publish a video to become viral?
- Can we make any inferences about viewer preferences based on the analysis of the video data?
- What are the implications for advertising and monetization?