Industry-standard automatic bat call classifiers produce high numbers of false positives for Pipistrellus echolocation calls, especially Pipistrellus pygmaeus. In this project, myself and data scientist Soffie Wisniewska will produce a machine learning model capable of high accuracy identification of Pipistrellus sp. calls. This model will be used to improve efficiency of bat call analysis by ecologists. We aim to build the model to allow human operator feedback such that accuracy increases with use. 

​

This project is ongoing. Updates will be posted here, and annotated code will be available on my Github repository.