[Paper Review] Valuating User Data in a Human-Centric Data Economy

Marius Paraschiv, Nikolaos Laoutaris|arXiv (Cornell University)|Aug 27, 2019

Privacy, Security, and Data Protection17 references4 citations

TL;DR

This paper proposes a fair, transparent method to value and compensate users in a human-centric data economy using the Shapley value from cooperative game theory. It applies this framework to a movie recommendation system, demonstrating that user contributions—measured by rating behavior—vary significantly in value, with popular or timely ratings contributing more to revenue, and validating the approach with scalable approximation algorithms that outperform naive methods.

ABSTRACT

The idea of paying people for their data is increasingly seen as a promising direction for resolving privacy debates, improving the quality of online data, and even offering an alternative to labor-based compensation in a future dominated by automation and self-operating machines. In this paper we demonstrate how a Human-Centric Data Economy would compensate the users of an online streaming service. We borrow the notion of the Shapley value from cooperative game theory to define what a fair compensation for each user should be for movie scores offered to the recommender system of the service. Since determining the Shapley value exactly is computationally inefficient in the general case, we derive faster alternatives using clustering, dimensionality reduction, and partial information. We apply our algorithms to a movie recommendation data set and demonstrate that different users may have a vastly different value for the service. We also analyze the reasons that some movie ratings may be more valuable than others and discuss the consequences for compensating users fairly.

Motivation & Objective

• To address the imbalance in the current data economy where users generate value but receive no direct compensation.
• To develop a fair and transparent method for assigning monetary value to user data contributions in online services.
• To apply cooperative game theory—specifically the Shapley value—to quantify individual user contributions to a recommender system’s performance and revenue.
• To design scalable approximation algorithms that estimate user value efficiently without sacrificing fairness or interpretability.
• To enable users to verify their compensation through a transparent, privacy-preserving accounting layer.

Proposed method

• Uses the Shapley value from cooperative game theory as a theoretical foundation for fair value allocation among users based on their data contributions.
• Applies the Shapley value to measure each user’s marginal contribution to the performance of a movie recommendation system, using movie ratings as input.
• Employs clustering and dimensionality reduction techniques to reduce computational complexity in large-scale user sets.
• Uses partial information and sampling-based approximations to estimate Shapley values efficiently, avoiding the exponential cost of exact computation.
• Maps estimated user values to financial compensation, assuming a fixed total revenue share to be redistributed to users.
• Proposes a transparent meta-data layer to allow users to audit and verify their compensation based on their behavioral data.

Experimental results

Research questions

• RQ1 How can user data contributions be fairly valued in a human-centric data economy?
• RQ2 What makes certain user ratings more valuable than others in a recommendation system?
• RQ3 How can the computationally expensive Shapley value be approximated efficiently while preserving fairness and interpretability?
• RQ4 What are the key behavioral patterns that increase a user’s contribution to a service’s revenue?
• RQ5 How can users be fairly compensated based on their actual data contributions without compromising privacy?

Key findings

• Users who rate popular or newly released movies contribute significantly more to the recommendation system’s performance and thus have higher data value.
• The Shapley value provides a theoretically fair benchmark for user compensation, but exact computation is infeasible for large-scale systems.
• Approximation algorithms based on clustering and dimensionality reduction produce results that align with common-sense assessments of user value.
• The value of a user’s data is not uniform—some ratings (e.g., for trending films) are more valuable than others due to their impact on system accuracy and revenue.
• The framework enables transparent, auditable compensation, increasing user trust and participation in data-driven platforms.
• The proposed method supports generalization across different data types and use cases, such as traffic data or social media, once key metrics are identified.