Resumo

Unstructured data plays a key role in the consumer decision-making process. Advanced techniques for linguistic analysis allow for extracting meaning from the content provided by social media influencers. In this paper, we identify the key dimensions of video content on YouTube using text analytics methods, such as topic modeling and sentiment analysis, in a comprehensive study of main content topics on YouTube, social media influencers, and comparisons of user digital engagement.

The study's goal is to investigate what the most popular content and the valence-associated social media influencers discuss on YouTube. Then, we propose a unified framework for (1) extracting the latent content-related topics from social media influencers' channels on YouTube; (2) ascertaining the labels, valence, and heterogeneity of those dimensions; and (3) using those dimensions for strategy analysis considering digital engagement measures.

We use social media influencers (SMI) as content creators, digital engagement, and video features on YouTube as a theoretical framework.

The method used comprises collecting audio transcriptions from videos, after executing a text preprocessing, performing a topic modeling stage, the Latent Dirichlet Allocation algorithm (LDA), and text analysis. The study collected data on the number of views, likes, dislikes, comments, topics content, and other video post characteristics, from 34,563 videos posted on YouTube among 103 different YouTubers channels. Data collection was conducted by application programming interfaces (APIs) via Python programming language, and data persistence was performed using a MySQL relational database.

We named the 50 topics such that it reflects the topic of discussion being evaluated and then grouped some of them based on semantic similarities, video content, and channel participation. Resulting in 19 different content labels: Beauty, Culture & Entertainment, Decoration, Organization & DIY, Education, Economics, Entrepreneurship & Business, Entertainment/general, Family, Fashion/Lifestyle, Gaming, Gardening, Gastronomy, Health & healthy lifestyle, Military, People, Behavior & Lifestyle, Pets & Animals, Politics, Economy & News, Sports, Tech, and Travel, learnings & curiosities.

The study highlights the top 3 content categories with greater digital engagement among influencers: 'Family', 'Entertainment/general', and 'Culture and Entertainment'. The sentiment analysis shows that content about 'Beauty', 'Gastronomy' and 'Economics, Entrepreneurship and Business' are those with the highest proportional positive valence. And, 'Politics, Economy and News', 'Entertainment/general', and 'Gaming' with high percentages of negative valence. The study results can help identify the influencer's audience's interests, preferences, and potential content gaps in the SMI's channel.

Aleti et al. (2019). Tweeting with the stars: Automated text analysis of the effect of celebrity social media communications on consumer word of mouth. Journal of Interactive Marketing, 48, 17-32. Hughes et al. (2019). Driving brand engagement through online social influencers: An empirical investigation of sponsored blogging campaigns. Journal of Marketing, 83(5), 78-96. Ladhari et al. (2020). YouTube vloggers’ popularity and influence: The roles of homophily, emotional attachment, and expertise. Journal of Retailing and Consumer Services, 54, 102027.