A Mapping Lens for Estimating Data Value

Public data is a key resource for society. The digital revolution and a renewed focus on evidence-based policymaking have made estimating the value of public data resources more urgent than ever. However, conventional methods to assess their value are still underdeveloped and generally fail to capture their broader impact. We introduce a novel approach to assess the value of data, anchored in its potential to improve decision-making. This approach likens data to a map, in that it provides an (imperfect) representation of a conceptual landscape that guides how people navigate the landscape. We define the value of data as the divergence in the quality of decisions taken with this map as compared with the closest alternative, and argue for the use of natural experiments to estimate this value. We demonstrate the use of this approach with four concrete examples drawn from our past research. Crucially, our method can inform how federal agencies decide to collect, maintain, and publish their critical data assets amid growing privacy concerns.

---