Developers of retail applications can benefit from FiftyOne’s ability to process large amounts of data and query/filter for specific instances. For example, in a few clicks or lines of code, you can get a subset of your data that includes shoplifting instances of a specific store, which you can then quickly verify for false positives or curate into a training dataset. The FiftyOne Brain provides various methods to aid in dataset/model analysis, one of which being the ability to find visually similar images which can be useful to train product recommendation systems.