Generating Futures from Text. Scenario Development using Text Mining

In foresight, scenarios illustrate probable, plausible and possible future developments and serve as a framework for strategic planning and decision-making. Scenarios try to draw holistic images considering various aspects of today's world. Currently, their development is complex and time consuming. For example, at the beginning of the scenario development process in most cases literature needs to be screened in order to develop a fundament of influences for the scenario stories. This article focuses on the initial phase of scenario preparation. It investigates if and to which extent scenario development may be optimized by text mining, an approach to automatically process texts and aggregate their content (scientific publications and reports in this case). The attempt is made to apply appropriate methods to summarize the topic and gain indicators for driving aspects. In order to draw a comparison, two solutions are realized, one based on a commercial software and one own implementation. To test and evaluate the procedure, two case studies are conducted. As the results show, the delimitation and structuring of the scenario field is supported and input for discussing the influences is delivered.