New publication: Measuring bias in cultural data using Soviet newsreel materials

Image: provided by the authors.

Cultural data is inherently biased, as it reflects interpretations of what is considered important or interesting. For example, foreign news coverage does not report all events occurring worldwide; rather, it emphasizes those deemed relevant to the target audience. Examining such biases is crucial because it helps uncover the cultural, geographical, and temporal perspectives embedded in the data—perspectives that are often difficult to detect. In this paper, we propose a method for measuring bias and demonstrate its application to a collection of Soviet newsreel materials.

Affiliation / context: This article belongs to a series of papers emerging from the multidisciplinary research collaboration at the Cultural Data Analytics Lab (CUDAN), Tallinn University.

Read the publication