Docker for Data Science: Building Scalable and Extensible Data ... - Joshua Cook - Google Books
Learn Docker "infrastructure as code" technology to define a system for performing standard but non-trivial data tasks on medium- to large-scale data sets, using Jupyter as the master controller.It is not uncommon for a real-world data set to fail to be easily managed. The set may not fit well into access memory or may require prohibitively long processing. These are significant challenges to skilled software engineers and they can render the standard Jupyter system unusable. As a solution to this problem, Docker for Data Science proposes using Docker. You will learn how to use existing pre-compiled public images created by the major open-source technologies—Python, Jupyter, Postgres—as well as using the Dockerfile to extend these images to suit your specific purposes. The Docker-Compose technology is examined and you will learn how it can be used to build a linked system with Python churning data behind the scenes and Jupyter managing these background tasks. Best practices in using existing images are explored as well as developing your own images to deploy state-of-the-art machine learning and optimization algorithms.What You'll Learn Master interactive development using the Jupyter platformRun and build Docker containers from scratch and from publicly available open-source imagesWrite infrastructure as code using the docker-compose tool and its docker-compose.yml file typeDeploy a multi-service data science application across a cloud-based systemWho This Book Is ForData scientists, machine learning engineers, artificial intelligence researchers, Kagglers, and software developers