A Linguistic Approach to Categorical Color Assignment for Data Visualization.

Area of Science:

• Data Visualization
• Computational Linguistics
• Human-Computer Interaction

Background:

• Data visualizations benefit from semantically meaningful color associations.
• Existing methods may not fully leverage linguistic properties of data terms for color generation.

Purpose of the Study:

• To explore the use of linguistic information for generating semantically meaningful colors in data visualization.
• To develop a robust method for identifying and assigning colors based on term-color associations.

Main Methods:

• Utilized co-occurrence measures from Google n-grams to define a 'colorability' score for terms.
• Employed semantic analysis and Google Images for color retrieval.
• Leveraged WordNet for symbolic relationships to assign identity colors.
• Applied k-means clustering for creating visually distinct color palettes, with options for predefined palettes.

Main Results:

• Established a quantitative measure ('colorability') for term-color association strength.
• Successfully retrieved representative colors for data categories using linguistic and visual data.
• Generated visually distinct color palettes through clustering, adaptable to predefined color constraints.

Conclusions:

• Linguistic information can effectively guide the generation of semantically meaningful colors for data visualization.
• The proposed methods provide a scalable approach for creating contextually relevant and visually appealing color palettes.
• This technique enhances data interpretation by aligning visual elements with conceptual understanding.