Inspiration
I was inspired to do this project by a couple of my friends, who post good YouTube videos, but don't ever go trending.
What it does
This project analyzes the best publish time, title, and category for YouTube videos to have to go trending the most. It also finds the relationship between views and number of comments, as well as between views and number of likes.
How we built it
Created in PyCharm with Python modules including numpy, scipy, matplotlib, seaborn, and pandas.
Challenges we ran into
Finding the mode of a large dataset, as well as calculating correlation coefficients and other statistical values turned out to be more difficult than expected.
Accomplishments that we're proud of
Proud to have found solutions to my problems by just searching online and finding the best module for the job. Especially with the correlation coefficient and p-value calculations, I dreaded making the calculations myself until I discovered SciPy to do it for me inside its own module.
What we learned
I learned how to use SciPy a little bit. I also learned more about saving files with matplotlib and the different kinds of seaborn plots while experimenting with different visual representations for the data.
What's next for An Analysis on Trending YouTube Videos
Next, I want to try one-hot encoding video titles somehow so that I can find a way to implement machine learning in a decision tree to spontaneously find and differentiate good and bad titles for videos.
Log in or sign up for Devpost to join the conversation.