Barrier islands provide resources and ecological services that are integral to economic and environmental interests, such as protection of coastal infrastructure and providing habitat for wildlife. Over time, barrier islands may become eroded and experience land loss, which require management actions to restore island integrity. The process of restoring barrier islands can create new habitats but also alter existing habitats, which can impact the organisms depending on coastal habitats, such as the Endangered Species Act-listed (ESA) Piping Plover (Charadrius melodus) and Red Knot (Calidris canutus). Consequently, we aimed to understand the abundance and behavioral responses from a suite of shorebird species, including American Oystercatcher (Haematopus palliatus), Red Knot, Piping Plover, Snowy Plover (Charadrius nivosus), and Wilson’s Plover (Charadrius wilsonia), to restoration and habitat factors at two restored sites in coastal Louisiana, Whiskey Island and Caminada Headland, from 2012 to 2020. We combined observational data from birds with numerous geospatial predictor variables into a machine learning approach to develop response plots to investigate how birds use coastal habitat with respect to restoration activities at Whiskey Island and Caminada Headland, Louisiana from 2012 to 2020. For each species, we quantified the abundance and behavior (breeding, foraging, leisure) of each individual, which served as response variables in boosted regression tree models, and determined the importance of predictor variables, such as restoration phase, habitat class, inundation class, distance to water, and other remotely or field collected variables from Whiskey Island and Caminada Headland. We ran at least three models for each species, which included non-breeding abundance with a Poisson distribution, foraging behavior with a binomial distribution, and leisure behavior with a binomial distribution. For American Oystercatcher (only occurring in reasonable sample sizes at Whiskey Island) and Wilson’s Plover, we also ran models for breeding abundance with a Poisson distribution and breeding behavior with a binomial distribution.