Enabling Interactivity on Displays of Multivariate Time Series and Longitudinal Data
Temporal data is information measured in the context of time. This contextual structure provides components that need to be explored to understand the data and that can form the basis of interactions applied to the plots. In multivariate time series we expect to see temporal dependence, long term and seasonal trends and cross-correlations. In longitudinal data we also expect within and between subject dependence. Time series and longitudinal data, although analyzed differently, are often plotted using similar displays. We provide a taxonomy of interactions on plots that can enable exploring temporal components of these data types, and describe how to build these interactions using data transformations. Because temporal data is often accompanied other types of data we also describe how to link the temporal plots with other displays of data. The ideas are conceptualized into a data pipeline for temporal data, and implemented into the R package cranvas. This package provides many different types of interactive graphics that can be used together to explore data or diagnose a model fit.