With our Spinnaker Sit-Down interview, we give you VIP access to industry influencers, thought leaders, and movers-and-shakers who share their unique insights and perspectives on industry issues, trends, and questions. This month we sat down with data scientist Dr. Howard Friedman.
Dr. Friedman has the unique background of having worked in the private sector, public sector, and academia. He works full-time for the United Nations where he focuses on statistics and health economics, and is known for his role as a lead statistical modeler on a number of key UN projects. He is also an adjunct professor at Columbia University where he teaches data analytics. In the private sector, he has led data teams in private equity, banking, and retail as well as leading health and economics research teams for the healthcare industry.
Dr. Friedman has authored or co-authored dozens of articles and book chapters in the areas of applied statistics, health economics, and politics. His book “Measure of a Nation,” published in 2012, focuses on how to improve America by first comparing its performance to 13 competitive industrial nations, then identifying the best practices found throughout the world.
SPINNAKER: How did you get involved with data and analytics?
FRIEDMAN: I was always good at science and math, and always enjoyed them. When I was pursuing my PhD in biomedical engineering at Johns Hopkins University, I took advanced math and statistics classes, and they fascinated me. I fell in love with what we could do with data and was lucky enough to have an amazing statistics professor. I later learned that data acquisition and doing lab work were not for me. But figuring out what data means — for science, business, health — what information was telling me, and what it could be used for, that’s what I wanted to do.
SPINNAKER: You started your career at Capital One. Tell us about that.
FRIEDMAN: Capital One was a great fit for me, because they had an appreciation for the skill of “information-based strategy” — as far as I know, they coined the phrase. They hired people with quant skills not just to capture data but to use it and apply it to the real world — to risk modeling, to fraud identification, to forecast corporate loss, to predict the chance a customer would respond to a solicitation. When I was there, they were really ahead of the curve in terms of being a core data-driven company. They were a leader in minimizing risk losses for the company, and in emerging subprime customers. In fact, that’s where I met Spinnaker’s Shawn Sweeney and Brett Ludden, who did outstanding work in risk and analytics for Capital One.
SPINNAKER: What’s most exciting about your field?
FRIEDMAN: I really get excited about the ability to connect different data sources together and create a richer understanding, to find out who and what’s driving things at the moment.
SPINNAKER: Tell us about what the field of data and analytics looks like right now.
FRIEDMAN: Modeling methods are much more sophisticated than they were 20 years ago. The methods, software, and infrastructure are much better, the processing power is much better, and the volume has massively exploded. Today you can analyze more deeply, and avenues are opening up because of these enhanced capabilities. The key is figuring out the information I need to help me achieve what I want.
With capacity and interest at an all-time high, the potential pitfall is understanding. To really leverage the power of data and analytics, you need deep knowledge and subject matter expertise to succeed. Experience is a substantial advantage.
SPINNAKER: What are some of today’s challenges in the field?
FRIEDMAN: One of the challenges is the integrity of the analysis itself, understanding the biases built into the data that are potentially enhanced in your model. It’s critical to figure out how to dedupe the biases and remove data errors and other issues that could hinder model development.
Data privacy and data ethics are also challenges, understanding where you can and can’t go in the area of data analytics. How could or should you apply or operationalize solutions? What’s appropriate and not appropriate? What are customer expectations? That’s where finding strong partners from a broad range of backgrounds within the business is critical, to guide ethics in understanding algorithms, to define legal limitations about what you can and can’t do, and to bring to the table knowledge of the softer side of customer expectations, all so you can determine what is practically effective.
SPINNAKER: “AI” is a hot topic right now. What can you tell us about AI?
FRIEDMAN: AI is super-hot right now. But the phrase AI is used in many, many ways, with a vague definition that creates confusion. In general, when something is in the buzzword phase, it’s used by a lot of people who are not clear what exactly the term means. AI is great technology and there are great opportunities to leverage it across the public and private sector. That said, AI has had a long history of peaks and troughs in terms of excitement. There’s a risk of AI being overpromised and underdelivering. Take image prediction. A computer can tell the difference between a cat, a dog, and a stop light in an image. But that’s not generalized intelligence. That’s not the same as how humans learn and transfer knowledge from one problem to another. Rather, that is more like curve fitting. A huge challenge is transfer knowledge, where the capacity that has been developed to solve one problem is immediately applicable to another problem. AI has already had some incredible accomplishments in image and speech recognition as well as other modeling applications but there is a risk that it is being oversold right now.
Python-based software is relatively easy for the user to learn, and is incredibly powerful. The positive and negative of powerful software that is easy to use is that users may not necessarily understand the assumptions underneath the models. Users can create impressive models and be successful very quickly even if they’re not formally trained in statistics, data analytics or data science. That works fine until assumptions are created that blow up the model. So it’s a mixed blessing. Users feel ownership and it involves a broader community. But the risk is that subject matter expertise and sometimes statistical expertise is not as appreciated. More and more users can develop good models, with sometimes less recognition for what’s missing. They need to understand basic statistics. They would benefit from a guide into the modeling, like a subject matter expert with years of experience who knows how to avoid the cliffs. They need to know how to use it once it’s built. For this, business experience becomes a tremendous value.
SPINNAKER: What is your secret to success?
FRIEDMAN: First, I don’t think I’m all that successful. That said, I think that four key elements in success are: (1) Effort —hard work can get you very far regardless of where you start. (2) Talent — natural talent. Find your niche and then you shine. (3) Connections — in the real world, connections matter, and you can build them. And (4) Luck — luck means a lot!
More About Dr. Friedman:
Native: New Yorker
Languages: English, French, Spanish, Chinese
Childhood career dream: didn’t think about it
Life lesson: my parents taught me the importance of community and connecting with people, which has probably helped me become a better communicator than perhaps the typical data nerd
Gets best ideas: when I’m doing something else, a dot gets connected
Passion: travel – I have been to over 75 countries. While I like going to popular destinations, I find it much more exciting to get off the beaten path and have been to countries including Papua New Guinea, Burkina Faso, Zimbabwe and Mongolia. (see photos of places he has visited and worked on his website)
First job: in my uncle’s pharmacy doing whatever he said to do from deliveries, cash register, stock boy, and fetching him cheeseburgers — I learned the value of a strong work ethic
Quote: “Learning is the only thing the mind never exhausts, never fears, and never regrets,” Leonardo da Vinci
Best compliment: after a corporate event, someone told me I was the only person with a big title that ever helps clean up — I believe in being part of a team, not above a team