# Virality Prediction and Community Structure in Social Networks

**Authors:** Lilian Weng, Filippo Menczer, Yong-Yeol Ahn  
**arXiv:** [1306.0158](https://arxiv.org/abs/1306.0158)  
**Date:** 2013-06-02  
**Summary page:** https://www.summarizepaper.com/en/paper/1306.0158v2/

## Key points

- Network structure influences the diffusion of information
- Impact of network structure on diffusion varies depending on the type of contagion
- Complex contagions are influenced by social reinforcement and homophily
- Highly clustered communities enhance the spread of complex contagions within them
- Diffusion across communities is hindered for complex contagions
- Memes and behaviors can be classified as complex contagions
- Most memes exhibit characteristics of complex contagions
- Some viral memes spread across numerous communities, similar to diseases
- Predicting meme popularity can be done by quantifying its early spreading pattern in terms of community concentration
- The more communities a meme permeates, the more likely it is to become viral
- Data about community structure can be translated into predictive knowledge about information spread
- Connection between network structure and virality prediction has implications for computational social science, social media analytics, and marketing applications

## Summary

The study titled "Virality Prediction and Community Structure in Social Networks" explores how network structure influences the diffusion of information. Previous research has indicated that the impact of network structure on diffusion varies depending on the type of contagion. Unlike simple contagions such as infectious diseases, complex contagions are influenced by social reinforcement and homophily. This means that within highly clustered communities, the spread of complex contagions is enhanced while diffusion across communities is hindered. The authors investigate whether memes and behaviors can be classified as complex contagions. They find that most memes indeed exhibit characteristics of complex contagions. However, they also identify a few viral memes that spread across numerous communities, similar to diseases. To predict the future popularity of a meme, the researchers propose quantifying its early spreading pattern in terms of community concentration. The more communities a meme permeates, the more likely it is to become viral. The study presents a practical method for translating data about community structure into predictive knowledge about which information will spread widely. This connection between network structure and virality prediction has significant implications for computational social science, social media analytics, and marketing applications. Overall, this research sheds light on how different types of contagions interact with network structures and provides insights into predicting the spread of information in social networks.