Bar graphs are routinely used in academic works, official reports, and mass media. Prior studies have focused on the comprehension of numerical information in bar graph design but have largely ignored the semantic information representation. Actually, along with the escalating need to convey semantic information beyond numerical data, unconventional bar graphs emerge and catch increasing eyes, highlighting the necessity of unlocking semantic information representation in bar graph design. In this paper, we attempt to address these gaps through examining the impact of three visual channels-color, shape, and orientation-on viewers' comprehension of semantic information. Drawing from prior research, we formulate a series of research hypotheses and conduct two experiments. Results show that by evoking sensorimotor experiences, conceptually relevant colors and shapes of bars facilitate the representation of semantic information. This facilitation is more pronounced in conveying concrete concepts than abstract concepts. Similarly, by evoking emotional experiences, colors and orientation aligned with the affective valence of concepts aid the representation of semantic information, with a more noticeable enhancement in conveying abstract concepts compared to concrete concepts. Additionally, we find that shape-embellished bars somewhat hinder the judgment of specific numerical values. These findings provide a renewed perspective on how semantic information is represented in bar graphs, offering valuable practical guidance for scientifically representing semantic information.

Urban landmarks play an important role as spatial references in spatial cognition, navigation, map design and urban planning. However, the current landmark extraction methods do not consider the imbalance between the landmark and non-landmarknon-landmark samples in a dataset, so the extraction results are biased toward the class with the majority of sample data, resulting in poor classification performance for the class with the fewest sample data. This study introduces a random forest (RF) classifier combined with cost-sensitive learning to extract urban landmarks automatically from a basic spatial database. First, the optimal feature set is determined according to the importance of features. Next, a cost-sensitive RF algorithm is applied to extract landmarks, which determines the misclassification cost according to the class distribution, and each decision tree is weighted by the classification results. The method has good performance, with a recall and area under the ROC curve (AUC) greater than 90%, and the model is also applicable to small sample sets, which can reduce the cost of manual labor.