This work is driven by the need for a conceptual framework to tackle temporal analyses in a data-centric workflow. Most research on temporal data focuses on modeling. Corresponding software requires very stringently formatted data, but does little in providing guidelines or tools for wrangling raw data into the required format. This has led to ad-hoc, and once-off solutions, which this research repairs.

There are three original contributions for the temporal data analysis in this research. They are grounded in the spirit of exploratory data analysis for time-indexed data. The first contribution (Chapter 2) is a new technique for visualizing data using a calendar layout. It is most useful when the data relates to daily human activity, and patterns can be explored in relation to people’s schedules. The second contribution (Chapter 3) is a new data abstraction which streamlines transformation, visualization, and modeling in temporal contexts. This “tsibble” object is a data infrastructure forming the foundation of temporal data pipelines. The third contribution (Chapter 4) is to equip analysts with exploratory and explanatory tools for discovering missing patterns in time, and thus facilitating decisions on appropriate imputation methods for further downstream analysis.