Interview with Feyzi Bagirov, Faculty of Analytics, Harrisburg University - Speaker at Global Predictive Analytics Conf - March Posted on : Mar 07 - 2017

We feature speakers at Global Predictive Analytics Conference March 27 - 29, 2017 to catch up and find out what he or she is working on now and what's coming next. This week we're talking to Feyzi Bagirov, Faculty of Analytics, Harrisburg University.

Interview with Feyzi Bagirov

1. Tell us about yourself and your background.
In 2006 I was the second employee in a newly-established Customer & Business Intelligence department at Dassault Systemes. After getting my MBA in 2012, I consulted University of Maryland University College during the creation of Master of Science in Data Analytics and also taught at Southern New Hampshire University. 

In 2015, I've built a Bachelor of Science in Data Science program for Becker College. While at Becker I've also built a model, predicting student enrollment and developed a Game Analytics class for their flagship Game Design program. 

2.  What have you been working on recently?
Currently myself and two other professors are working on a 529 project - an educational project, which launches the pilot with  universities in several African countries. Analyzing data that will come from the pilot is going to be a major step of the project. I also teach Principles of Analytics at Harrisburg University of Science and Technology and consulting a San-Francisco based startup Metadata.io on Data Science. 

3. Tell me about the right tool you used recently to solve customer problem?
In 2015, a research showed, that 23.45% of analysts used both R and Python concurrently; that was up from 20% in 2014. 

All analytical applications have/can do statistical algorithms that 99% of the analysts need(R might have an edge when it comes to the number of packages/libraries, but others are catching up quickly).

I am tool-agnostic. My philosophy is that the more tools you know, the better. Confining yourself to a specific tool will not let you take advantage of the best features other tools have to offer. 

I might want to use R for explorative analysis, and then Python for web scraping, data cleanup and modeling. I will then either switch back to R or use Tableau for visualizations. 

When it comes to one thing that is going to solve most of the customer problems, 99 times out of a 100 it is data integration/data cleanup. Cleaning up the dirty data can take up to 3/4 of the analytical project. Any existing or new data wrangling library is helpful. 

4. Where are we now today in terms of the state of Predictive Analytics, and where do you think we’ll go over the next five years?
From what I've seen in other countries, US organizations are at the forefront of leveraging the data analytics. Many companies successfully implementing predictive and prescriptive analytical models. However, there are two problems that I see stalling the process in organizations - data and human capital. 

Many companies, especially where data is not the core competency (like hospitals, etc.) still have little grasp of what data they have, where it is stored and how to get it. If they can get it, they do not know what can/should they do with it. They are throwing it on a data analyst, without having any specific questions in mind that they want this data to answer. 

The other problem is human capital. As McKinsey predicted in 2011, by the next year (2018), the US alone will face a shortage of 140-190,000 people with deep analytical skills, as well as 1.5 million managers and analysts with the know-how to use the analysis of big data. Although since then many successful academic and vocational programs have been established (including boot camps), the situation with the human capital within the organizations is more complicated - people are used to and still using Excel, mostly working on descriptive analytics, and are afraid of math and coding.

On the other hand, Artificial Intelligence, and specifically Machine Learning, achieved the level, where algorithms can perform jobs that require certain level of intelligence - lawyers, doctors and accountants included. 

I think in the next 5 years, as we will see more jobs automated, we will also see a shift in education and training, shifting more towards STEM education and hands-on training. Similar shift happened after Henry Ford automated assembly line and started mass producing cars for consumers - a lot of people in the horse-breeding and carriage-riding industry had to re-train to become car mechanics and car drivers. 

We will definitely see more analytical innovations in the areas of cyber security, Internet-of-Things, digital marketing and healthcare. 

5. What are some of the best takeaways that the attendees can have from your "Acquisition Funnel for Higher Education" talk?
-Be realistic to narrow the scope of the predictive modeling work at the initial stage
-Evaluation of models should be in the process
-Building data structure is more important than methodology

6. What are the top 5 Predictive Analytics Use cases in enterprises?
1) Marketing analytics - advertising campaigns, retention and churn modeling, uplift for direct modeling, cross sales and recommendation engines. 

2) Using analytics to identify Fraud, Waste and Abuse - this goes more towards government, infrastructure, telecommunications, healthcare and other similar service industries. 

3) Cyber security analytics - 80% of cyber security crimes are behavior based. SAS released a Cyber Security Analytics product in 2015. 

4) Predictive Maintenance analytics - manufacturing, infrastructure, oil and gas, military, automotive, logistics and transportation industries will see the biggest value of this type of analytics. 

5) Natural Language Processing - mining through the massive texts of customer call center comments, potential candidates' resumes, identifying sentiment for the brand in the online community, etc.  

7. Which company do you think is winning the global Predictive Analytics race?
I would call out Deep Knowledge Ventures, a Hong Kong VC fund that appointed an algorithm (VITAL) to its board. 
In reality, the companies with the best Data Science/Analytics teams are positioned to make the biggest impact. Majority of them are, naturally, in the technology sector: Google, Microsoft, Facebook, Linkedin, Twitter, Amazon and many others have amazing data science teams. 

8. Any closing remarks
Utilizing organizational data is more than running descriptive dashboards. Getting into a predictive component quickly is important.