3.5 Discussion

Exploratory data analysis involves many iterations of finding and summarizing patterns. With temporal data available at finer scales, exploring time series has become overwhelming with so many possible granularities to explore. A common solution is to aggregate and look at the patterns across the usual granularities such as hour-of-day or day-of-week, but there is no way to know the “interesting” granularities a priori. A huge number of displays need to be analyzed or we might end up missing informative granularities. This work refines the search for informative granularities by identifying those for which the differences between the displayed distributions are greatest and rating them in order of importance of capturing maximum variation.

The significant granularities across different datasets (individuals/subjects) do not imply similar patterns across different datasets. They simply mean that maximum distributional differences are being captured across those granularities. A future direction of work is to be able to explore and compare many individuals/subjects together for similar patterns across significant granularities.