Recommending Views from Stacked Graph Logs

We developed a view recommender module to encourage and support creativity. Our module was built for the stacked graph visualization and is based on a first study on implicit interest indicators. In our system, dwell time on a view is considered as good user interest indicator. Based on our first study, we examined several dwell times thresholds for identifying interesting views and found that the most effective threshold was 11.5 seconds. Views satisfying this property were considered to calculate the dwell time mean and views above the mean were considered as good candidates for recommendation. Our goal is provide a tool to facilitate creativity by knowledge gathering, knowledge integration and ultimately, idea generation.