Youtube Analysis (Updating)
Proposal
Different kinds of video features and elements would enhance the performance of each youtube channel. We decided to analyze the data from YouTube regarding different video elements and features. The objective of this topic is to find useful video elements that can maximize performance from a business perspective.
Handling the scrolling page, retrieving all video links in each scroll and store them for later use
Loop through each scrolled page. Within this loop, scrape each page.
Delete Missing Value
Transform data Type
Remove Outliers
The number of views across the year
The peak is around Nov & Dec
The number of views vs the length of videos
5-15 mins and 15-30 mins have more views than others
The number of subscribers vs views
No Strong Correlation with .35(<.5) pairwise correlation
Over 75% of videos have a percentage of like of 3.78, which means every 100 views would only generate roughly 4 likes.
According to the standard between likes and videos, the ratio for measuring success would be 4 likes for every 100 views.
Companies should focus on quality instead of quantity
