Two Tier Analysis of Social Media Collaboration for Student Migration

Abstract

Global adoption of Social Media as the preferred medium for collaboration and information exchange is increasingly reshaping social realities and facilitating new research methodologies in various disciplines. Social Media applications are collecting a large amount of User-Generated Content (UGC) and web data that contains knowledge about novel approaches of global collaboration between people. We have done a detailed study of the factors that lead to student migration, as espoused by social scientists, and compared it with factors observed by analyzing over 10 million Twitter posts. Using the gravity model as our baseline, we built a novel methodology to identify the features and facts that twitter posts offer for studying human collaboration during migration. We leveraged methods from Natural Language Processing (NLP) to extract contents specific to migration from social media posts. We used topic modeling- Latent Dirichlet Allocation (LDA) to extract the topics from tweets and word embedding- Word to vector (W2V) to find the correlation and similarity between UGC and socioeconomics theories. In this paper, we present our methodology in detail, along with the results of our analysis.