A major challenge in ecology and evolutionary biology is to understand how biological patterns at one scale are generated by a multitude of processes operating at various scales. Two approaches are especially powerful at linking these processes to patterns: mathematical modeling and using a citizen science dataset. This dissertation uses these two approaches to detect unobservable biological processes from observable patterns in three separate studies: monarch butterfly population dynamics in eastern United States, microbiota population dynamics in fly gut, and mating strategy dynamics in hybridizing species pairs. First, the dissertation tests whether a continental population decline of the monarch butterfly is caused by the scarcity of milkweed, as the milkweed decline has been shown to locally impact monarch population. The study concludes that the milkweed scarcity is not the cause of the continental monarch decline. An observation made at a microscopic scale cannot be extrapolated to explain the pattern at a macroscopic scale. Second, the dissertation develops novel method to understand population dynamics of ingested bacteria from fecal time-series taken from its host. Application of this method to experiments using Drosophila shows that bacterial population is regulated over larger gut area in the host as the density of bacteria increases. Information on processes at a microscopic scale may be preserved at a macroscopic scale. Third, the dissertation develops a model to understand how unequal population sizes between hybridizing species pairs influence the evolution of mate choosiness in these species. An observation of greater choosiness in a smaller population has often been interpreted as evidence for reinforcement, but our results suggest that this interpretation is not valid. Some microscopic processes may dominate others to generate macroscopic patterns. Lastly, this dissertation highlights the importance of scale not only in basic science, but also in applications such as conservation and medicine.