Continuous Integration for Data Science · Pivotal Engineering Journal

This is a follow up post on Test-Driven Development for Data Science and API First for Data Science focusing on Continuous Integration. Last time we wrote about the importance of test-driven development for data science, especially in the context of what we call smart applications. There are many examples for smart applications, for example, Google's Inbox has a feature that is called Smart Reply which uses machine learning to suggest three possible answers to your incoming messages. Another instance of a smart app is Apple's Photo app on the iPhone. While taking a photo, it recognizes faces and places through machine learning and automatically relates them together so that searching for photos becomes much smarter.