[Paper Review] Social Browsing on Flickr

Kristina Lerman, Laurie A Jones|ArXiv.org|Dec 7, 2006

Recommender Systems and TechniquesComputer Science4 references111 citations

TL;DR

This paper investigates social browsing on Flickr, demonstrating that users primarily discover new images through their contacts' photo streams rather than tags or groups. Using analysis of user activity, the study finds that social networks—especially reverse contacts—strongly predict image popularity, with 75% of comments on high-quality images originating from a photographer's social network, highlighting social browsing as a key mechanism for content discovery and recommendation.

ABSTRACT

The new social media sites - blogs, wikis, del.icio.us and Flickr, among others - underscore the transformation of the Web to a participatory medium in which users are actively creating, evaluating and distributing information. The photo-sharing site Flickr, for example, allows users to upload photographs, view photos created by others, comment on those photos, etc. As is common to other social media sites, Flickr allows users to designate others as ``contacts'' and to track their activities in real time. The contacts (or friends) lists form the social network backbone of social media sites. We claim that these social networks facilitate new ways of interacting with information, e.g., through what we call social browsing. The contacts interface on Flickr enables users to see latest images submitted by their friends. Through an extensive analysis of Flickr data, we show that social browsing through the contacts' photo streams is one of the primary methods by which users find new images on Flickr. This finding has implications for creating personalized recommendation systems based on the user's declared contacts lists.

Motivation & Objective

• To understand how users discover new images on Flickr beyond traditional methods like tags and groups.
• To investigate the role of social networks—specifically contacts and reverse contacts—in shaping user browsing behavior.
• To evaluate whether social browsing is a dominant mode of content discovery compared to other discovery mechanisms on Flickr.
• To determine the extent to which social network size and structure predict image popularity and engagement metrics.
• To assess the feasibility of using social networks for personalized image recommendation systems.

Proposed method

• Collected and analyzed data on three image sets: Random (randomly selected), Apex (user-selected high-quality images), and Explore (Flickr’s algorithmically selected top images).
• Tracked engagement metrics: views, favorites, and comments on images across the three sets.
• Mapped user social networks by analyzing contact lists and reverse contacts (users who have the photographer as a contact).
• Correlated engagement metrics with social network size and structure, particularly the number of reverse contacts.
• Used statistical analysis to compare the influence of social networks, tags, and group submissions on image popularity.
• Examined comment sources to determine whether commenters were from the photographer’s social network (mutual, reverse, or strangers).

Experimental results

Research questions

• RQ1 To what extent does social browsing through contacts influence user discovery of new images on Flickr?
• RQ2 How do engagement metrics (views, favorites, comments) correlate with the size of a photographer’s social network?
• RQ3 What is the relative impact of social networks, tags, and group submissions on image popularity?
• RQ4 How do comment sources (mutual contacts, reverse contacts, strangers) differ across image sets (Random, Apex, Explore)?
• RQ5 Can social network structure predict whether an image will be featured on Flickr’s Explore page?

Key findings

• Social browsing through contacts is the dominant method for discovering new images on Flickr, with 75% of comments on high-quality images originating from the photographer’s social network.
• The number of reverse contacts (users who have the photographer as a contact) correlates most strongly with image engagement, including views and favorites.
• Images in the Apex and Explore sets show similar engagement patterns despite the Apex set being months old, indicating that the Interestingness algorithm effectively identifies high-quality photographers.
• Only 10% of Apex images were previously featured on the Explore page, suggesting limited overlap between user curation and algorithmic selection.
• Tags are less effective than social networks for content sharing, and group submissions play a minor role except for Random users, possibly due to weaker social networks.
• Despite public exposure, the size of a photographer’s social network remains the key factor in getting images featured on the Explore page.