Unlocking Social Insights with Computational Social Science: Big Data, Social Media, and NLP in Public Discourse

Academic Publication

In the digital age, social media is more than a platform for conversation—it’s a goldmine of social data. Every post, comment, and share creates digital traces that reveal how people interact, influence each other, and form opinions. Using computational social science methods, big data analytics, and natural language processing (NLP), researchers can uncover patterns in online discourse that traditional surveys often miss.

Mapping Public Opinion on Nutrition Labels in the EU

Figure 1. Network analysis. The circumference of the circle shows the total amount of tweet data. The circumference is divided into segments, each color corresponds to a country and each dot is associated with their conversation ID. The segments are connected by chords that illustrate the relationship within/between countries. Each tweet relationship is symbolized by a separate chord. (Original figure by Irawan et al., Frontiers in Nutrition (2022). Source: https://doi.org/10.3389/fnut.2022.846730)

Figure 2. Network analysis of dialogues. The Fruchterman–Reingold layout algorithm shows tweets belonging to dialogues and having similar conversation IDs. The dialogues were represented by colors, each corresponds to the different themes in which the dialogues were coded. The nodes (tweets) are connected by lines, representing their connections. (Original figure by Irawan et al., Frontiers in Nutrition (2022). Source: https://doi.org/10.3389/fnut.2022.846730)

In one of my studies, I analyzed Twitter conversations about front-of-pack nutrition labeling (FOPL) across the European Union. Using computational text analysis, sentiment analysis, and social network analysis, we explored not only the content of discussions but also who drives the conversation.

Rather than treating social media as a mere communication channel, this research views it as a complex social system. Certain actors scientists, institutions, or citizen advocates—play central roles, shaping how information spreads and how public sentiment is formed. The study highlights inequalities in influence and visibility, showing how network structures determine who is heard in online debates. (Source: https://doi.org/10.3389/fnut.2022.846730)

Listening to Conversations on Children Nutrition in Indonesia

Building on this approach, my more recent study currently under review examined social media discourse on child stunting in Indonesia. By applying social media listening, peak detection, NLP-based sentiment analysis, and engagement metrics, we identified critical moments of public attention and mapped how different actors shaped the narrative.

The findings reveal patterns of influence and attention: some voices dominate the conversation, while others remain on the periphery. These insights are crucial for public health communication and policy-making, helping to identify key influencers and understand how collective sentiment evolves over time. (Source: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4888137)

Why Computational Social Science Matters

Both studies demonstrate the power of computational social science and big data to transform social media data into actionable insights. By combining large-scale data collection, NLP, and network analysis, we can study:

• How online communities interact

• How influence and attention are distributed

• How social norms, expertise, and inequality are negotiated in digital environments

This approach provides a deeper, more nuanced understanding of public discourse than traditional methods alone.

Looking Ahead

As computational social science, NLP, and big data continue to advance, their applications in public health, policy, and social research are expanding. Analyzing social media at scale helps researchers, practitioners, and policymakers understand how people engage with information, how influence operates online, and how collective opinions form, enabling more effective, inclusive communication strategies.

Social media is not just a tool for conversation—it’s a window into the social structures of our digital world, and computational methods are key to unlocking its potential.