Spatio-temporal data are ubiquitous in ecology. As the volume of ecological data increases, they encompass long time periods on large spatial domains, frequently involving multiple species and variables. Developing methods to summarize these data while providing ecological interpretability is crucial for maximizing the insights they offer. Empirical Orthogonal Functions (EOFs) are a well-documented and widely used method for reducing the dimensionality of spatio-temporal data. First, introduced in the 1950s, EOFs essentially involve performing Principal Component Analysis on spatio-temporal data. Subsequently, a substantial body of literature has developed around EOFs, primarily in the fields of meteorology and climate studies. While a substantial body of research has applied EOFs in meteorology and climate science, their use in ecology is relatively new and remains overlooked, despite their huge potential. In this paper, we aim at presenting the basics of EOFs and introduce new related methods that address specific ecological questions. To illustrate their potential, we use two different datasets: (1) satellite data for the Normalized Difference Vegetation Index (NDVI) over France, and (2) spatio-temporal predictions from an integrated species distribution model that maps several fish species in the Bay of Biscay. Since EOFs do not incorporate ecological constraints, their results may not always be directly interpretable from an ecological perspective. In this paper, we develop a method that incorporates a temporal ecological variable into EOFs to improve their interpretability. Using satellite NDVI data, we demonstrate how EOFs results can be informed by including a steady seasonal variable, which helps to better represent a North-South gradient across France. Finally, EOFs can be a valuable tool for investigating community and ecosystem spatio-temporal dynamics, though this implies to move to a multivariate framework. We illustrate several ways to leverage EOFs for multivariate analysis and highlight shared spatial and temporal patterns among several spatio-temporal variables. Using multispecies fish data from the Bay of Biscay, we identify both common temporal trends and shared spatial patterns in reproduction timing and locations. A set of reproducible codes are provided in the manuscript to assist ecologists interested in applying these techniques in practice.