Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
102 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

The Anatomy of the Facebook Social Graph (1111.4503v1)

Published 18 Nov 2011 in cs.SI and physics.soc-ph

Abstract: We study the structure of the social graph of active Facebook users, the largest social network ever analyzed. We compute numerous features of the graph including the number of users and friendships, the degree distribution, path lengths, clustering, and mixing patterns. Our results center around three main observations. First, we characterize the global structure of the graph, determining that the social network is nearly fully connected, with 99.91% of individuals belonging to a single large connected component, and we confirm the "six degrees of separation" phenomenon on a global scale. Second, by studying the average local clustering coefficient and degeneracy of graph neighborhoods, we show that while the Facebook graph as a whole is clearly sparse, the graph neighborhoods of users contain surprisingly dense structure. Third, we characterize the assortativity patterns present in the graph by studying the basic demographic and network properties of users. We observe clear degree assortativity and characterize the extent to which "your friends have more friends than you". Furthermore, we observe a strong effect of age on friendship preferences as well as a globally modular community structure driven by nationality, but we do not find any strong gender homophily. We compare our results with those from smaller social networks and find mostly, but not entirely, agreement on common structural network characteristics.

User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (4)
  1. Johan Ugander (47 papers)
  2. Brian Karrer (41 papers)
  3. Lars Backstrom (6 papers)
  4. Cameron Marlow (3 papers)
Citations (984)

Summary

The Anatomy of the Facebook Social Graph

In "The Anatomy of the Facebook Social Graph," Ugander, Karrer, Backstrom, and Marlow present an extensive quantitative analysis of the structural properties of the social graph of active Facebook users in May 2011. With 721 million active users and 68.7 billion friendships, the paper administers an exceptional lens to scrutinize the largest social network analyzed to date.

Key Observations

Global Structure and Connectivity

The paper finds remarkable connectivity within the Facebook graph, noting that 99.91% of users belong to the largest connected component. This implies a nearly fully connected graph on a global scale. Importantly, the analysis supports the "six degrees of separation" phenomenon with an average pairwise distance of 4.7, indicating that most user pairs are connected by a short path.

Degree Distribution

Contrary to the often-assumed power-law distribution, the degree distribution of the Facebook social graph exhibits significant curvature on a log-log scale, implying deviations from simple power-law behavior. The degree distribution is highly skewed, with most users having fewer than 200 friends, while a smaller subset possesses degrees in the thousands, constrained by a 5000 friendship limit imposed by Facebook.

Clustering and Local Neighborhoods

The analysis of local clustering coefficients reveals that even within a globally sparse network, local neighborhoods exhibit substantial density. The average clustering coefficient falls with increasing degree, yet remains significantly higher than what has been observed in other social networks like MSN messenger. This paper highlights a distinctive structure in which users’ friend networks form tightly-knit communities.

Moreover, the concept of degeneracy further elaborates this density, revealing robust k-cores within user neighborhoods, with average degeneracy substantially higher than previously studied large networks. This indicates that even the sparse Facebook network possesses dense core communities among users, especially those with higher degrees.

Degree Assortativity

Positive degree assortativity characterizes the Facebook social graph; users with more friends are more likely to connect with other users who also have more friends. The nuanced exploration of the conditional probability p(kk)p(k'|k) affirms that higher-degree users are predominantly connected to other high-degree users. This assortative mixing extends to the phenomenon where individuals’ friends generally have more friends than the individuals themselves—an observation with significant implications on users' social perception.

Implications and Future Directions

Practical Implications:

  1. Algorithm Development: The insight that users' friend networks are highly clustered and contain dense subgraphs is pivotal for the design of efficient graph traversal and search algorithms. The local density can significantly impact BFS and other traversal algorithms’ operational complexity, necessitating optimization strategies that account for high clustering.
  2. Social Media Strategies: The observation that Facebook users are connected to a broad and densely linked global component is vital for understanding information dissemination, viral marketing, and content propagation strategies on social media platforms.

Theoretical Implications:

  1. Network Science: The discrepancies between Facebook’s degree distribution and the classic power-law model encourage the development of more nuanced models to better capture real-world social networks' complexity.
  2. Sociology and Behavioral Studies: The analysis of assortativity and mixing by demographics such as age and engagement offers a granular view of underlying social structures and peer influence, providing a basis for deeper sociological research into online behavior patterns.

Speculative Future Developments in AI:

  1. Enhanced Social Graph Analysis: Future AI models could incorporate these findings to enhance predictive analyses, such as identifying influential nodes or understanding community dynamics with greater accuracy.
  2. User Behavior Prediction: The correlation between user engagement and network position could inform AI systems designed to predict user behavior, enabling platforms to personalize user experiences and optimize content delivery based on inferred social connections.

In conclusion, "The Anatomy of the Facebook Social Graph" provides an exhaustive characterization of the largest social network, contributing critical insights into the global connectivity, local density, degree assortativity, and demographic mixing patterns within Facebook. The findings have substantial implications for both practical applications and theoretical advancements in social network analysis and algorithm design.

Youtube Logo Streamline Icon: https://streamlinehq.com