Interactive visual analysis promotes exploration of long-term ecological data

Long-term ecological data are crucial in helping ecologists understand ecosystem function and environmental change. Nevertheless, these kinds of data sets are difficult to analyze because they are usually large, multivariate, and spatiotemporal. Although existing analysis tools such as statistical methods and spreadsheet software permit rigorous tests of pre-conceived hypotheses and static charts for simple data exploration, they have limited capacity to provide an overview of the data and to enable ecologists to explore data iteratively, and interactively, before committing to statistical analysis. These issues hinder how ecologists gain knowledge and generate hypotheses from long-term data. We present Ecological Distributions and Trends Explorer (EcoDATE), a web-based, visual-analysis tool that facilitates exploratory analysis of long-term ecological data (i.e., generating hypotheses as opposed to confirming hypotheses). The tool, which is publicly available online, was created and refined through a user-centered design process in which our team of ecologists and visualization researchers collaborated closely. The results of our collaboration were (1) a set of visual representation and interaction techniques well suited to communicating distribution patterns and temporal trends in ecological data sets, and (2) an understanding of processes ecologists use to explore data and generate and test hypotheses. We present three case studies to demonstrate the utility of EcoDATE and the exploratory analysis processes using long-term data on cone production, stream chemistry, and forest structure collected as part of the H.J. Andrews Experimental Forest (HJA), Long Term Ecological Research (LTER), and US Forest Service Pacific Northwest Research Station programs. We also present results from a survey of 15 participants of a working group at the 2012 LTER All Scientists Meeting that showed that users appreciated the tool for its ease of use, holistic access to large data sets, and interactivity.