Apply Storytelling Techniques for Describing Time-Series Data
Narrative and storytelling have played an important role in communication. In this work, we demonstrate that the techniques of storytelling, and in particular raising attention to abnormality, can be used to add interestingness and memorability to descriptions of time-series data. A computational system has been developed for automatically generating descriptions for data graphs. The system identifies visual patterns in the graph, and treats the graph’s deviations from corresponding ideal patterns as abnormal events. It then uses storytelling templates to generate a graph description with the abnormal events highlighted.
KeywordsNarrative Graph description Natural language generation
- 1.Arnheim, R.: Art and visual perception. University of California Press (1974)Google Scholar
- 2.Bailey, T. Gatrell, A.: Interactive spatial data analysis. Longman Scientific & Technical Essex (1995)Google Scholar
- 5.Boylen, C., Eichler, L., Swinton, M., Nierzwicki-Bauer, S., Hannoun, I., Short, J.: The State of the Lake: Thirty Years of Water Quality Monitoring on Lake George (2014). https://fundforlakegeorge.org/StateoftheLake
- 6.Chatman, S.: Story and Discourse: Narrative Structure in Fiction and Film. Cornell University Press (1980)Google Scholar
- 16.Rubin, D.C.: Memory in oral traditions: The cognitive psychology of epic, ballads, and counting-out rhymes. Oxford University Press on Demand (1999)Google Scholar