5 Minutes with a Data Scientist: Gregory Piatetsky-Shapiro of KDnuggets
What's a data scientist do, anyway? What's the job like, what helps data scientists succeed, and where is this job headed? Gregory Piatetsky-Shapiro, Ph.D., is the president of KDnuggets. He is a well-known expert in data mining and data science and is the cofounder of KDD conferences and ACM SIGKDD professional organization. He was chief scientist at two start-ups. In this interview, he talks about what it means to be a data scientist.
UPSIDE: First, tell us about your job.
Piatetsky-Shapiro: My current job as president of KDnuggets bringing our readers (more than 300,000 unique monthly visitors to www.kdnuggets.com and 120,000 subscribers/followers via email and social networks) the interesting opinions, tutorials, software, etc. in the hot fields of big data, data science, and machine learning. There is so much happening, but our goal is to select just a few interesting stories each day.
If you could go back in time, what's the one thing you would tell yourself as a new analyst/data scientist?
Keep learning the best tools and languages (such as Python, R, and Scala) and learn more linear algebra and calculus -- essential knowledge for taking deep learning even deeper!
What's a personality trait you think people need to succeed at your job?
Curiosity, analytical inclination, and a willingness to question assumptions.
What's your biggest pet peeve (abused buzzword, overhyped idea, etc.) and why?
I really don't like the term "citizen data scientist," primarily because it implies that people without much training can do the work of a data scientist.
The big demand for data science skills can be met by either training more people in data science or by creating fully automated software, similar to the software in a plane's autopilot or in current self-driving cars.
What will not work is creating partly automated software and expecting people with little training or understanding of data to be able to use it in new situations.
Imagine a combination of an untrained pilot and an autopilot that works 90 percent of the time. The plane will fly fine for a while until an unpredictable situation arises, and the autopilot will signal a problem and revert control to the untrained pilot. Bad things will happen.
Similar logic is behind Google's decision to remove a steering wheel from its self-driving cars -- a driver who is relying on a car almost all the time is unlikely to react quickly and adequately in an emergency when the car cannot cope.
For more, see my full article on this subject: "The Mirage of a Citizen Data Scientist," and a related KDnuggets cartoon.
Where is data science headed in the next few years?
I think data science, like many other fields, is headed towards increased automation.
Only four years after an article by Thomas Davenport and DJ Patil in the Harvard Business Review proclaimed "Data Scientist: The Sexiest Job of the 21st Century" comes a recent post by the same Thomas Davenport: "Six Very Clear Signs that Your Job Is Due to Be Automated." This article included these criteria for jobs likely to be automated:
- It involves answering data-dependent questions
- It involves quantitative analysis
- It involves the creation of data-based narratives
All these criteria describe the job of a data scientist!
Moreover, a recent KDnuggets poll, "Data Scientists Automated and Unemployed by 2025?" found the majority of respondents thought that expert-level data science will be automated by 2025.
This will not eliminate all data scientist jobs, but it will affect many of the lower-skilled positions. Those data scientists who want to remain relevant need to focus on skills that are harder to automate, such as creativity, intuition, and human communication skills.
James E. Powell is the editorial director of TDWI, including the Business Intelligence Journal and Upside newsletter.