Ecological complex networks representing predator-prey interactions are crucial for analyzing the multilayered relationships, structure, function, and dynamics of ecological systems. However, constructing comprehensive networks has been challenging due to the resource intensive process of capturing and validating the interactions across diverse taxa and broad spatio-temporal scales, often resulting in models that are biased towards well-observed regions and organisms. This paper presents a novel approach that leverages scientifically validated, crowd-sourced observations from an iNaturalist project to construct and analyze a predator-prey network encompassing thousands of interactions and taxonomic groups ranging across multiple spatio-temporal dimensions. This project is the first to model a real-time scalable empirical network using validated citizen science observations that uncovers species importance, investigates energy flow patterns, identifies clusters and motifs, and analyzes robustness and stability to perturbations using data collected from across continents. This work demonstrates the practicality of combining citizen science with ecological network analysis and candidly assesses limitations of using crowdsourced data for holistic ecosystem research.