[Paper Review] Social Networks and Social Information Filtering on Digg

Kristina Lerman|ArXiv.org|Dec 7, 2006

Recommender Systems and TechniquesComputer Science3 references66 citations

TL;DR

This paper investigates social filtering on Digg, a social news aggregator, demonstrating that users are significantly more likely to engage with stories liked or submitted by their friends, validating social filtering as an effective information recommendation method. The study reveals that social networks amplify visibility, but also risk 'tyranny of the minority' where a small group of interconnected users dominates front-page content, prompting algorithmic changes to improve diversity.

ABSTRACT

The new social media sites -- blogs, wikis, Flickr and Digg, among others -- underscore the transformation of the Web to a participatory medium in which users are actively creating, evaluating and distributing information. Digg is a social news aggregator which allows users to submit links to, vote on and discuss news stories. Each day Digg selects a handful of stories to feature on its front page. Rather than rely on the opinion of a few editors, Digg aggregates opinions of thousands of its users to decide which stories to promote to the front page. Digg users can designate other users as ``friends'' and easily track friends' activities: what new stories they submitted, commented on or read. The friends interface acts as a \emph{social filtering} system, recommending to user stories his or her friends liked or found interesting. By tracking the votes received by newly submitted stories over time, we showed that social filtering is an effective information filtering approach. Specifically, we showed that (a) users tend to like stories submitted by friends and (b) users tend to like stories their friends read and liked. As a byproduct of social filtering, social networks also play a role in promoting stories to Digg's front page, potentially leading to ``tyranny of the minority'' situation where a disproportionate number of front page stories comes from the same small group of interconnected users. Despite this, social filtering is a promising new technology that can be used to personalize and tailor information to individual users: for example, through personal front pages.

Motivation & Objective

• To evaluate the effectiveness of social filtering in recommending news stories on Digg by analyzing user behavior.
• To investigate how social networks influence story visibility and promotion to Digg’s front page.
• To examine unintended consequences such as 'tyranny of the minority' where a small group of active users disproportionately shapes front-page content.
• To compare Digg’s social filtering approach with Reddit’s collaborative filtering, assessing their relative effectiveness.
• To assess the impact of algorithmic changes on user diversity and front-page representation.

Proposed method

• Tracking the voting patterns of stories over time to analyze user engagement with content shared by friends.
• Analyzing user social networks by identifying mutual friends and tracking friend activity (e.g., submissions, diggs, comments).
• Visualizing the network of mutual friends among top-ranked users to detect clusters and influence propagation.
• Measuring the maximum number of diggs attained by front-page stories as a function of the submitter’s user rank.
• Comparing pre- and post-algorithm-change data (November 2006) to evaluate the impact of reduced voting diversity on front-page diversity.
• Using statistical analysis to correlate user rank with story popularity and front-page promotion rates.

Experimental results

Research questions

• RQ1 Do users tend to like stories submitted by their friends on Digg?
• RQ2 Do users show increased engagement with stories their friends have read and liked?
• RQ3 To what extent do social networks contribute to the promotion of stories to Digg’s front page?
• RQ4 Does the dominance of a small, interconnected group of users lead to 'tyranny of the minority' in front-page curation?
• RQ5 How effective is social filtering compared to collaborative filtering, as seen in Reddit’s approach?

Key findings

• Users are significantly more likely to dig stories submitted by their friends, confirming the effectiveness of social filtering.
• Users also show higher engagement with stories their friends have liked, indicating that social signals propagate interest.
• A single cluster of 30 top-ranked users with mutual connections dominates front-page visibility, contributing to 'tyranny of the minority'.
• After Digg changed its promotion algorithm to prioritize voting diversity, the average number of front-page stories per user dropped from 1.6 to 1.2, indicating improved diversity.
• The new algorithm successfully reduced the dominance of top users, though long-term effects on social network formation remain uncertain.
• Social filtering enables personalized front pages based on friends’ activity, offering a solution to the 'tyranny of the majority' in global front-page curation.