A Metric-Based Evaluation and Methodological Comparison of Infographics in Cartography

This study evaluates four assessment approaches—questionnaire-based evaluation, Quantitative Content Analysis (QCA), visual summary, and machine learning-based image analysis—for quantifying infographic style in cartographic products. These methods are compared against the IGV (Infographic eValuation) metric, a novel method proposed by the authors. All approaches were applied to a reference dataset comprising 32 images, equally divided between maps and infographics. The questionnaire survey (n = 139) employed a bipolar scale (−5 to +5) to measure perceived affiliation with cartographic or infographic categories; aggregated scores were subsequently normalised to a [0, 1] interval to ensure cross-method comparability. QCA was operationalised through 20 paired codes structured into four principal indicators (linking visual style, visualisation complexity, visual appeal and cartographic accuracy), generating composite scores. The visual summary approach was quantified by transforming predefined domain-based diagram segments into numerical values. Machine learning experiments implemented in Orange (Image Analytics) included hierarchical clustering using cosine distance and supervised classification via logistic regression trained on 50 labelled images. The article details the procedural steps of this evaluation, the results obtained, and a comprehensive statistical analysis. Furthermore, it assesses the suitability of each method for evaluating infographics and identifying specific elements within cartographic products.