Skip to content
O'Reilly home
Learning Path

Machine Learning for Healthcare Using Python, TensorFlow, and R

Instructor Nicole Tache
Time to complete: 7h 27m

Published byO'Reilly Media, Inc.

CreatedJanuary 2018

What is this learning path about, and why is it important?

Artificial intelligence (AI) and machine learning are having dramatic impacts on an ever-broadening array of industries. In few other fields, however, is that impact expanding as rapidly or with more promise than in healthcare. The marriage of technology and medical science offers great potential for assisting medical professionals and advancing patient care in yet unimaginable ways, all while drastically helping to cut the cost of healthcare. But many people affected by this field or interested in joining the burgeoning health-tech industry know very little about what is possible with current technology, and the rate at which the technology is progressing is exponential, making it that much more difficult to get a handle on it all. Where can you go to examine the state of the art and learn what’s coming down the pike?

In this learning path, your host, software engineer Aileen Nielsen, examines what can be done in healthcare now, which tools are most effective, and which concepts are most crucial to producing valuable applications in the near future. You’ll do this from the perspective of three of the most widely used machine learning tools: Google TensorFlow and the Python and R languages. You’ll first gain a conceptual overview of mature machine learning techniques useful for health-related applications, including data pipelines to turn unstructured data into useful inputs, unsupervised learning methods to search for novel insights from large health datasets, and forecasting and classification techniques for health metrics. Next, Aileen introduces you to three specific types of unsupervised learning, cluster analysis, association analysis, and principal components analysis, as applied to health datasets, both at the individual and population levels. You’ll then explore why linear methods remain a powerful and sophisticated way to think about data for prediction, causal analysis, and optimization in health tech. Finally, you will learn the basics of neural networks, including training neural networks with both image-based and unstructured healthcare data.

What you’ll learn—and how you can apply it

  • Understand the basics of putting together a health-tech data pipeline from raw datasets
  • The data challenges inherent in many scenarios within healthcare applications, from medical records to the quantified self
  • The three broad domains of machine learning as applied to healthcare: unsupervised learning, linear methods, and deep learning
  • Understand how to make causal inferences in health data using R and Python
  • Survey a range of current neural network applications in healthcare using Python and TensorFlow

This learning path is for you because…

  • You're a product manager or technical lead at a health-related organization, and you want to know whether your existing applications can expand into something more sophisticated, and whether it’s worth it
  • You're a data scientist or data engineer, and you want to work in health tech AI
  • You need a basic understanding of upcoming challenges and opportunities in the industry to know where you can apply your skills
  • You're a data analyst who has worked in traditional data roles outside of healthcare, but you want a better understanding of what neural networks and deep learning are all about
  • You're a software engineer who wants your healthcare-oriented organization to use the hottest new machine learning technologies


  • You should have an interest in how computers can affect health care
  • You should have a basic understanding of mathematics and statistics as well as programming in Python and R
  • You should be familiar with machine learning (i.e., how training works and how to prepare data for training a model)
  • Although not required, a familiarity with linear algebra would be helpful

Materials or downloads needed in advance: None