Archive reconstruction: This is an independently written research summary for a historical URL. It is not the original article, and this site is not affiliated with the original author.
Research on the re-identification risks of publishing social-network structures without names or other direct identifiers.
Historical context
Online social networks can be represented as graphs of people and their connections. Removing usernames from such a graph does not necessarily remove unique structural patterns; overlapping data from another network may enable matching.
Why the topic matters
The historical study is frequently discussed because it made clear that social relationships themselves can be sensitive information. Even seemingly anonymous datasets may reveal who knows whom, and that can expose associations a person intended to keep separate.
Reading the material today
For data stewards, the enduring lesson is to evaluate likely auxiliary information and downstream uses before treating a graph release as anonymous. A useful risk review examines more than the presence or absence of obvious personal fields.
Why Anonymous Social Networks Can Still Reveal Personal Identity
Social-network data contains more than usernames, profile photographs and personal descriptions. The connections between people form a network structure that can itself be revealing. Removing names from such a dataset does not necessarily remove the distinctive patterns of friendship, membership and association contained within it.
Research into social-network deanonymization showed how auxiliary information could be compared with supposedly anonymous network data. Overlapping connections may help an observer match individuals across datasets, potentially exposing relationships the original data publisher intended to keep private.
This concern is particularly relevant when social information is shared or repurposed. Advertising partners, online communities and social applications may process activity and relationship data for analytics or personalization. Data that seems harmless on its own can become more sensitive when combined with information from another service.
The Electronic Frontier Foundation raised related concerns in a 2012 discussion of online dating privacy, drawing attention to how sharing ostensibly anonymized social information might still compromise individuals. The underlying research helps explain why privacy reviews should consider network structure as well as direct personal identifiers.
The 33 Bits of Entropy research archive places these findings within a wider history of research on online anonymity, behavioral traces and data re-identification. Across these topics, the recurring question is not simply whether names were removed, but whether remaining characteristics could identify someone when combined with information available elsewhere.
For researchers and data stewards, the practical lesson is to assess likely external datasets, structural uniqueness and downstream uses before describing a social graph as anonymous.
Editorial context (2026): This is a new research summary for a restored historical permalink, not the original 2009 article.
Historical source: Read the original material preserved at 33bits.wordpress.com.
