Machine Learning Times
Machine Learning Times
EXCLUSIVE HIGHLIGHTS
SHARE THIS:

3 weeks ago
Wise Practitioner – Predictive Analytics Interview Series: Jennifer Lewis Priestley at Kennesaw State University

 

By: Eric Siegel, Founder, Predictive Analytics World for Business

In anticipation of her upcoming keynote presentation at Predictive Analytics World for Business Las Vegas, May 31-June 4, 2020, we asked Jennifer Lewis Priestley, Professor of Applied Statistics and Data Science at Kennesaw State University, a few questions about their deployment of predictive analytics. Catch a glimpse of her keynote presentation, How Leading Enterprises Leverage Universities to Boost Analytical Innovation and Tap Talent, and see what’s in store at the PAW Business conference in Las Vegas.

Q: As an academic leader in analytics, what are the biggest trends in how academia is approaching predictive analytics?

A: Analytics is still a very young academic discipline; the newest academic programs already look very different from the first “wave” of programs that were launched 10 years ago.  As I go around the country, the most prevalent themes in conversations amongst analytics program managers include (1) the rising importance of formally teaching ethics (2) the integration of experiential learning – meaning the role of capstone courses and applied projects and (3) the recognition of communication and “soft skills” as having just as much importance as “technical skills”.

Q: I want to follow up on the first point – tell me more about ethics and how that is being approached.

A: Its so important.  We have so many examples from facial recognition, breast cancer screening, hiring algorithms that are so blatantly biased against different demographics…we almost cant go a week without reading another story related to biased models/algorithms. These results are unintentional – no one sets out to develop a racist or misogynistic model.  Data scientists dont intend to do harm – the problem is that they don’t understand the harm they can do.  My opinion is that one of the biggest contributors to the ethical issues of “biased algorithms” is when people who are engaged in model development dont understand some of the basics of mathematics and statistics – how do you know to check for bias if you have never been taught what statistical bias is? much less the process of validation?  I know that Statistics is substantively less sexy than “predictive analytics”, “machine learning” and “data science”.  But many of the ethical issues that we are seeing today could have been addressed prior to operationalization with some Stats 101.

In academia, any social research we engage in must be screened by an Institutional Review Board – these Boards consider research through the lens of three concepts – justice, beneficence and consent.  Historically, computational disciplines like computer science and statistics have had very little engagement with IRBs.  Since most analytical initiatives either use human-generated data as inputs…or the final product (the algorithm or the model) will impact humans…we may want to start considering analytics and “data” the same way we consider “human subjects” and require these computational studies to go through and IRB review.

Q: What is happening in terms of teaching predictive analytics that the average practitioner with limited connection to academia might not be aware of?

A: Analytics/Data Science is treated differently at different universities.  Unlike well established disciplines like Marketing, which is universally housed in the business school, analytics/data science does not have a common academic home.  Some universities place the discipline in the College of Computing…or the College of Engineering…or the College of Mathematics/Statistics…or in some interdisciplinary unit that integrates multiple disciplines.  The students graduating from the analytics programs from each of these could be a great addition to your organization – but to be clear – they will have different skills and will likely approach problem solving from different orientations. 

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A: Ask yourself the question…how many .edu email addresses are in your contact list.

—————————–

Don’t miss Jennifer’s keynote presentation, How Leading Enterprises Leverage Universities to Boost Analytical Innovation and Tap Talent, at PAW Business on Wednesday, June 3, 2020 from 1:15 to 2:00 PM. Click here to register for attendance.

By: Eric Siegel, Founder, Predictive Analytics World for Business

One thought on “Wise Practitioner – Predictive Analytics Interview Series: Jennifer Lewis Priestley at Kennesaw State University

  1. Pingback: Wise Practitioner – Predictive Analytics Interview Series: Jennifer Lewis Priestley at Kennesaw State University - Machine Learning Times : Rlogger

Leave a Reply

Pin It on Pinterest

Share This