ABSTRACT
Text can be enhanced with extrinsic data – facts about words from other contexts and sources. SparkWords encode categoric, ordered, or quantitative data into words set out in text layouts such as prose, lists, or tables. These SparkWords can then be embedded in explanatory text, superimposed over underlying charts and maps, and cross-referenced to related visualizations and data comics by using the same formats in both the words and the visualizations. Formats in SparkWords can apply to subsets of characters in a word or codes, for example, to express values at ordered levels in a hierarchical code, or quantitative outcomes of each game in a series. These formats can be used in numeric tables to facilitate rapid identification of orders of magnitude; or in textual tables to add additional data to words. Examples include natural language generation of paragraphs describing a dataset; and an NLP example tagging word sets in dialogue to highlight repeated words, catchphrases, and logical inversions.
