IEEE Computer Graphics and Applications, 35, 3 (2015) 10-14
A growing number of visual analytic and statistical software tools are being built to deal with temporal event sequences. These tools often have difficulty in dealing with two problems: the volume of records (the number of records may grow to hundreds of millions, making it difficult to load or apply operations to the data) and the variety of patterns (longs records are unique and the variety makes it difficult to see global patterns such as relationships, clusters or gaps, as well as to identify errors or anomalies). We propose a taxonomy of analytic focusing strategies for temporal event sequences, based on our experience working with dozens of case studies using EventFlow. Most strategies can lead to reduction in volume and variety simultaneously.