Data Mining of Twitter Retweets: A Visual and Practical Representation

The task of classifying data observations based on a set of independent variables is one that has been studied extensively in both machine learning and traditional data analysis. Methods like linear and logistic regression, as well as newer methods such as Naive Bayes, XG-Boost, and others, were all created to complete the task of data classification. As social media activity continues to generate data from users around the world, new methods are being developed to include previously underutilized information. In the case of social media platforms like Twitter and Facebook, interactions like sharing and replying can be used to draw directed edges between users, each of which is represented as a node in a graph network. In this report, it is shown how using graph data to supplement a more traditional model is beneficial to classification performance. The visual representations of such interaction networks are also explored.