Sentiment Analysis of Marijuana Content via Facebook Emoji-Based Reactions

Utilizing emojis to understand sentiments of online social media posts has received much attention recently. In this work, we perform an emoji-based sentiment analysis and classifier to gain insights into users' emotional reactions to marijuana-related posts on Facebook. We collected about 15,000 posts and 14 million users' reactions from the High Times Magazine Facebook page, a top ranking marijuana-related content provider with a longstanding monthly publication for advocation of cannabis legalization. We developed R scripts and utilized the Google Cloud Prediction API to estimate the sentiment of the posts from posted texts and emoji-based reactions. Our analysis revealed that “LIKE” and “LOVE” are the most frequently used reactions, and they are strongly correlated by a correlation coefficient of 0.82. Our analysis also showed that the correlation between number of comment and reactions are similar across all emoji types, except “SAD”. Interestingly, we found that words with negative sentiment occurred much more often than words with positive sentiment. Particularly, “weed” was used more than 800 times while that of “love” was used 310 times. Our study can be utilized as an alternative tool for revealing insights into users' opinions towards marijuana that may potentially benefit public health surveillance applications.