Talk on Anonymizing Longitudinal Panel Data at Statistical Days 2025 in Geneva

At the Swiss Statistics Meeting (Statistical Days 2025) in Geneva, Oscar Thees delivered a presentation addressing “Anonymizing Longitudinal Panel Data.” The talk examined distinctive privacy concerns inherent in panel datasets, where the accumulation of quasi-identifiers across time makes individuals’ trajectories more unique and therefore more identifiable.

The presentation noted that modifications to sensitive information can expose individuals even when certain details are concealed. A central challenge involves maintaining analytical utility while safeguarding privacy, since trajectories must remain logically consistent within and across variables.

The speaker illustrated these complexities through concrete examples and presented potential remedies, aiming to inform decision-makers at statistical agencies and commercial organizations about the significance of properly tackling these challenges.