In this repository, we sought to better understand the factors that led to channel success on YouTube, specifically, the number of subscribers a channel has. We selected a dataset with information about over 1 million channels on YouTube, containing attributes such as channel name, description, keywords, subscriber count, views, join date, country, and more. By utilizing these features along with some engineered features, we hope to find the most important factors that drive channel growth on YouTube.
- Size: 1,095,242 rows, 876.84 MB
- Link: Kaggle Dataset
Subscriber_Prediction.pdf: A presentation that outlines the motivation and key findings from our modelingYouTube_Subscriber_Prediction.ipynb: Jupyter notebook containing all data cleaning and modelling
Please run our notebook YouTube_Subscriber_Prediction.ipynb top to bottom
William Qi, Edmund Doerksen, Jessica Ling