Visualization using stars and the pitfalls of nonlinear relationships

Abstract

The article focuses on the issue of data visualization using stars, which are commonly used to evaluate the quality of various services, products, or phenomena. The author points out that converting numerical data into stars is not a trivial task and can easily lead to misleading interpretations. Using a specific example of visualising the effectiveness of COVID-19 vaccines, he demonstrates how the choice of transformation function (e.g., linear, logarithmic, or based on order) significantly impacts the resulting impression of the data. The text emphasizes that none of the presented visualizations is unambiguously correct or ideal—each has its advantages and pitfalls. The author warns that visualization using stars (and other simplified displays) is easily misused for manipulative purposes and recommends caution when interpreting such graphs. The article concludes with a discussion of the nature of the relationship between antibody levels and the degree of protection against infection and encourages deeper reflection on such simplifications.