Similarity-Driven Visual-Interactive Prediction of Movie Ratings and Box Office Results
Feeras Al-Masoudi, Daniel Seebacher, Mario Schreiner, Manuel Stein, Christian Rohrdantz, Fabian Fischer, Svenja Simon, Tobias Schreck, Daniel A. Keim · 2013
We present an approach developed in course of the VAST 2013 Mini Challenge: Visualize the Box Office. We follow a similarity-driven methodology to predict ratings and box office results based on historic data. An array of interactive visualizations allow ana-lysts to explore structured and unstructured data, activate their do-main background knowledge, and come up with predictions as a weighted sum of historically observed figures. We describe the workflow, our developed system, present results obtained during the Challenge execution, and discuss our method in light of exten-sion possibilities.