KDnuggets Home » News » 2015 » Mar » Opinions, Interviews, Reports » Interview: Satyam Priyadarshy, Halliburton on Unlocking Success for Big Data Projects ( 15:n10 )

Interview: Satyam Priyadarshy, Halliburton on Unlocking Success for Big Data Projects

We discuss Predictive Analytics in Oil & Gas industry, Big Data analytics, key drivers of success,common reasons of failure, trends, advice, and more.

satyam-priyadarshyDr. Satyam Priyadarshy is a Scientist, Technologist, Author, Coach, Faculty, Speaker, and is a pioneer in the fields of data science, big data, analytics, and emerging technologies. He is currently chief data scientist at Halliburton’s Land-mark PSL, leading expansion of integrated workflow capabilities and other data development initiatives.

Dr. Priyadarshy has appeared as speaker at several international conferences, and has written, co-authored, presented and published numerous research papers in peer-reviewed journals and magazines. He has held various leadership positions in AOL, Network Solutions, Acxiom Corporation prior to joining Halliburton. He was recently named to The Financial Times’ list of potential board candidates with emerging technology and analytics expertise. Dr. Priyadarshy is an adjunct faculty at Georgetown University. He is currently senior fellow at the International Cyber Security Center at George Mason University. He is advisory board member at multiple organizations including Big Data Summit, Virginia Tech’s MBA Board, etc.

Dr. Priyadarshy holds a Ph.D. from Indian Institute of Technology in Bombay, and an MBA from Virginia Tech.

First part of interview

Here is second and last part of my interview with him:

Anmol Rajpurohit: Q5. How do you assess the current maturity level of Predictive Analytics in Oil and Gas in-dustry? What major developments do you expect in the next 2-3 years?

oil-and-gasDr. Satyam Priyadarshy: Predictive analytics is well matured in many aspect of oil and gas industry. Note that oil and gas industry is upstream, mid-stream and downstream. Predictive analytics more mature in downstream oil and gas compared to upstream, because upstream data has higher degree of complexity, as mentioned earlier.

AR: Q6. How do you differentiate between traditional analytics (KPI focused) and Big Data analytics?

big-data-analyticsSP: In simple terms, I define KPIs (Key performance indicators) as essential to operate the business. It does not lead to any innovation or additional value creation for the enterprise. With Big Data analytics, not only one gets KPIs, but is constantly looking for outlier events, and creating new patterns from the every growing and increasingly connecting data sets.

AR: Q7. What are the key drivers of success in today's Big Data projects? What are the most common reasons behind failure? success-drivers

SP: The explosion of TAPS (Technologies, Applications, Products and Solutions) is one of the key driver for success in Big Data projects. For the data that one’s enterprise has, it is possible that various combinations of these agile TAPS will work to find new patterns and discoveries. One does not have to restrict to a fixed number of these TAPS to gain value.

The reasons for failure are few:
  1. The cultural and data silos within the organization
  2. The lack of knowledge and patience at the leadership level - Big Data analytics is about finding answers and new patterns and then asking the question why it is there. A large number of leaders are trained in traditional way of business intelligence, asking for answers for a question they have
  3. Agility in adapting emerging technologies

AR: Q8. Which of the current trends in Big Data arena are of great interest to you?

SP: Some of areas of interest from value creation point of view are:

  1. Real-time predictive models using streaming big data technologies
  2. The in-browser highly interactive information and scientific visualizations using Javascript and newer technologies on large volumes of data.

From a total cost of ownership point of view, a much cheaper storage that supports data-model free stor-age of massive amounts of historical raw data for easy access would be welcome technology. As the raw data size grows, the cost of traditional Hadoop and other platforms will become significant and not feasible for many businesses.

AR: Q9. What is the best advice you have got in your career?

build-your-careerSP: The advice from the cosmic powers is that you should leverage your infinite potential that is bestowed upon you, and have faith in your parent’s upbringing to achieve the success you want. Unless you want to build a career for yourself, no one will build your career.

AR: Q10. What key qualities do you look for when interviewing for Data Science related positions on your team?

entrepreneurialSP: The qualities that I look for in a person - if the person is entrepreneurial in nature, it is contrary to large corporation recruitment, but only such people can leverage their computer science and analytics skill, because they need to look for new patterns from the ‘data’ on a regular basis. They have to find things that are not obvious. This also requires that people have patience to deal with failures. As I say, finding gold in gold mine is a dirty and extremely high effort job, it takes typically 28 tons of raw ore to find 0.04 ounce of gold. Big Data analytics is about the same, high value for very small amount of gold.

AR: Q11. On a personal note, are there any good books that you have been reading lately, and would like to recommend?

SP: The books that inspire me originate from the land of Bharat, like the Hanuman Chalisa because it tells that a single entity can be successful in many areas, if they work sincerely for it.