My latest research interests have been in using computational methods to understand society, and in gaining a better understanding of the online platforms that are at the fabric of our every day connections. 

Sanyoura, L. E., & Anderson, A. (2022). Quantifying the Creator Economy: A Large-Scale Analysis of Patreon. Proceedings of the International AAAI Conference on Web and Social Media, 16(1), 829-840. 

Abstract In recent years, the “creator economy” has emerged as a disruptive force in creative industries. Independent creators can now reach large and diverse audiences through online platforms, and membership platforms have emerged to connect these creators with fans who are willing to financially support them. However, the structure and dynamics of how membership platforms function on a large scale remain poorly understood. In this work, we develop an analysis framework for the study of membership platforms and apply it to the complete set of Patreon pledges exceeding $2 billion since its inception in 2013 until the end of 2020. We analyze Patreon activity through three perspectives: patrons (demand), creators (supply), and the platform as a whole. We find several important phenomena that help explain how membership platforms operate. Patrons who pledge to a narrow set of creators are more loyal, but churn off the platform more often. High-earning creators attract large audiences, but these audiences are less likely to pledge to other creators. Over its history, Patreon diversified into many topics and launched higher-earning creators over time. Our analysis framework and results shed light on the functioning of membership platforms and have implications for the creator economy

​El Sanyoura, L. and Xu, Y. Gender convergence in the expressions of love: A computational analysis of lyrics. In Proceedings of the 42nd Annual Meeting of the Cognitive Science Society.

Abstract Love is a central theme in modern music, but do women and men differ in their expressions of love? Results from empirical studies on gender differences in love attitudes have evolved from showing consistent differences to more similarities over time and witnessed gender convergence in relationship expectations, housework responsibilities, and sexual attitudes. Independently, pop culture studies have shown how music can be used as a contextual artifact whose lyrics can reflect a culture's changing psychological processes and ideologies. We combine these two research areas to explore whether the gender convergence reported in psychological studies is mirrored in love songs. Using a corpus of lyrics and song metadata from 1960 to 2009, we present a computational analysis of the lexical distribution of lyrics across genre, gender and time. We show that love songs between vocalists who are men vs. women have become significantly more similar in their lyrical expressions of love.

​ Leiyu, J. El Sanyoura, L.  Xu, Y. How nouns surface as verbs: Inference and generation of word class conversion. In Proceedings of the 42nd Annual Meeting of the Cognitive Science Society.

Abstract Word class conversion refers to the extended use of a word from one grammatical class to another without overt morphological marking. Noun-to-verb conversion, or denominalization, is one form of word class conversion studied extensively in the literature. Previous work has suggested that novel denominal verb usages are comprehensible if the listener can compute the intended meaning based on shared knowledge with the speaker. However, no existing work has explored the computational mechanism under this proposal. We propose a frame-semantic generative model, Noun2Verb, that supports the inference and generation of novel denominal verb usages via semi-supervised learning. We evaluate this framework in a dataset of denominal verbs drawn from adults and children against a state-of-the-art model from natural language processing. Our results show that Noun2Verb aligns better with human interpretation and bridges the gap between machines and humans in lexical innovation.