Intentions in the Coordinated Generation of Graphics and Text from Tabular Data

Abstract

To use graphics efficiently in an automatic report generation system, one has to model messages and how they go from the writer (intention) to the reader (interpretation). This paper describes PostGraphe, a system which generates a report integrating graphics and text from a set of writer's intentions. The system is given the data in tabular form as might be found in a spreadsheet; also input is a declaration of the types of values in the columns of the table. The user then indicates the intentions to be conveyed in the graphics (e.g., compare two variables or show the evolution of a set of variables) and the system generates a report in L A T E X with the appropriate PostScript graphic files. PostGraphe uses the same information to generate the accompanying text that helps the reader to focus on the important points of the graphics. We also describe how these ideas have been embedded to create a new Chart Wizard for Microsoft Excel.

COSMOROE: a cross-media relations framework for modelling multimedia dialectics

“…These findings are only implicitly correlated with the interaction relations identified in the studies. Furthermore, the same studies have also looked at crossmodal references [1,2,21,23], without correlating these Table 1 Image-language relations in AI studies Work WIP system [1] Multimedia argumentation [23] Postgraph [15,16,20] Modalities 2D drawing-text 2D information graphics-captions 2D information graphics-captions findings with the interaction relation sets built either. While undertaken for dealing with very specific and diverse discourse genres, applications and modality types, the studies seem to share a-more or less-common glossary for expressing image-language relations.…”

“…We used a weight of 1.00 for all cases of full agreement, a weight of 0.00 for all cases of disagreement on the main relation, a weight of 0.50 for cases of subtype disagreement (non-applicable in our case) and a weight of 0.75 for cases of sub-subtype disagreement. 20 The weights actually determine the contribution of each category to the final agreement score. The weighted kappa reached 0.8851.…”

“…First of all, the task itself, i.e. to analyse the semantic relations among media for forming a message, is demanding, in terms of cognitive 20 For details on the calculation of kappa and of the weighted kappa cf. [17].…”

AutoBrief: an experimental system for the automatic generation of briefings in integrated text and information graphics

“…One early system, Postgraphe, creates a simple graph and caption based upon information from the user about his or her intentions, e.g. to emphasize a rising trend between 1980 and 1990 in the values of the price attribute in the user's data set (Fasciano and Lapalme, 1999). In Postgraphe design heuristics are used to design the graphic based upon the type of intention specified by the user (e.g.…”

The persuasive phase of visualization

“…PostGraphe [5,6] is another system that automatically generates both the graph and its explanatory captions. The graphical output it produces also varies depending on user intentions -a format is chosen from a fixed set based on considerations for intentions, data types, and expressiveness.…”