OpMAP is a tool that facilitates sense-making for large-scale debates. It contains several major components.
First, the debate is reconstructed, i.e. its arguments are analyzed and formalized. An illustrative reconstruction of the Veggie-debate is presented here.
Second, the logical analysis of the debate is used to construct a structured survey. Complex opinions on the debate’s topic can be elicited by means of this survey.
Third, the collected opinion data is translated into a so-called opinion graph. The nodes of this graph are the opinions held by the participants of the survey, and the weighted edges specify how closely related these opinions are. We use degree of mutual coherence to calculate how related two opinions are.
Then, the resulting weighted graph is analyzed by applying clustering and filtering methods.
The initial map of the debate is then constructed as described here.
Finally, OpMAP allows for dynamic updates of the map: As a new data record is obtained through the survey, the opinion graph, the clustering and the visualization are updated as described here.
To summarize, the steps to generate a dynamic opinion map are: