Perhaps they're used to the Scrapy APIs and tools? I agree that an HTTP requests library would have been sufficient but maybe it was just easier for them to use the framework they're used to.