Data science projects, by their very nature, take a long time to realize a return on investment. In particular, it is hard to accurately measure the success of projects that involve making long-term predictions. As noted in an earlier chapter, for departments to advance the cause of data science, it is essential for them to show early successes. In general, projects that a) are short term; b) have a measurable outcome; and c) can benefit and will be used a wide range of users are some of the key factors that help to establish credibility and ensure success of data science related projects.

An example of such a project is Arterys, a cloud-based company that developed a deep learning algorithm in late 2016 that could assess the flow of blood to the heart in a fraction of the time compared to conventional scanners. It met all the key criteria of success. The benefits were obvious almost immediately, the algorithm provided a directly measurable outcome since you could compare the results with those from scanners and it was useful to a broad range of users, namely, patients. Further, the topic was very intuitive.

Anyone who has ever had a blood test can intuitively relate to measuring bloodflow. The benefits of such a product are also quite obvious. Being able to get results in a fraction of the time could help save lives. In November 2016, the FDA approved the algorithm. It was a monumental achievement.