Back To Projects

Analysis of Trending YouTube Videos: Finding Patterns in Viral Content

Vincent P.

As the digital world continues to grow, content creators frequently have trouble building a community and producing videos that will interest their audience. Especially as these people look toward the internet for both recreational and monetary reasons, finding out techniques to build a community is important in today’s age. This paper analyzes the issues of video performance, revealing the patterns of what makes a video successful and viral. By training different models and testing different datasets, we were able to find the correlation between the potential chances of popularity and the video’s content. Using the most accurate model, the Random Forest model, content creators can see whether or not they are likely to do well based on patterns found in trending videos.


As the digital world continues to grow, content creators frequently have trouble building a community and producing videos that will interest their audience. Especially as these people look toward the internet for both recreational and monetary reasons, finding out techniques to build a community is important in today’s age. This paper analyzes the issues of video performance, revealing the patterns of what makes a video successful and viral. By training different models and testing different datasets, we were able to find the correlation between the potential chances of popularity and the video’s content. Using the most accurate model, the Random Forest model, content creators can see whether or not they are likely to do well based on patterns found in trending videos.

Explore More!

Vincent P.
Amanda Wang
MS in Computer Science, Computer Science and Business Analytics MIT Alum

Related Projects

Approaches to fraud detection on credit card transactions using artificial intelligence methods

In this paper, we study the problem of detecting fraudulent credit card transactions. We select the most relevant features using a heuristic approach, and fit three different model classes to a simulated dataset: Logistic Regression, Random Forests and Gradient Boosting Machines. We find that hyperparameter tuning has a big impact on the precision and recall of our classifiers. We also find that of the three classes, Gradient Boosting Machines were the best-performing model class, achieving 83% precision and 64% recall on unseen data.
Aryaman R.
Mentored by Yuan Lee
Predicting International Commodity Prices, Production, and Land Usage Towards Reducing Agricultural Emissions

Our long-term goal for this research project was to analyze food insecurity and determine if we could predict the emissions and pricing of different food commodities.
Sahir G.
Mentored by Landon Butler
Predicting Fake Job Listings from Real Ones using Machine Learning Models

The main way this problem is going to be solved is through the creation of models that can predict whether these job listings are real or fake.
Ronit M.
Mentored by Hassan Azmat