Date of Award
Department or Program
Department of Computer Science
The ability to successfully predict virality on Twitter holds great potential as a resource for Twitter influencers, enabling the development of more sophisticated strategies for audience engagement, audience monetization, and information sharing. To our knowledge, focusing exclusively on tweets posted by influencers is a novel context for studying Twitter virality. We find, among feature categories traditionally considered in the literature, that combining categories covering a range of information performs better than models only incorporating individual feature categories. Moreover, our general predictive model, encompassing a range of feature categories, achieves a prediction accuracy of 68% for influencer virality. We also investigate the role of influencer audiences in predicting virality, a topic we believe to be understudied in the literature. We suspect that incorporating audience information will allow us to better discriminate between virality classes, thus leading to better predictions. We pursue two different approaches, resulting in 10 different predictive models that leverage influencer audience information in addition to traditional feature categories. Both of our attempts to incorporate audience information plateau at an accuracy of approximately 61%, roughly a 7% decrease in performance compared to our general predictive model. We conclude that we are unable to find experimental evidence to support our claim that incorporating influencer audience information will improve virality predictions. Nonetheless, the performance of our general model holds promise for the deployment of a tool that allows influencers to reap the benefits of virality prediction. As stronger performance from the underlying model would make this tool more useful in practice to influencers, improving the predictive performance of our general model is a cornerstone of future work.
Han, Danah K., "Predicting Influencer Virality on Twitter" (2020). Dartmouth College Undergraduate Theses. 155.