Research topic for Stats Phd, working as Business Analytics Manager

Dear all,

I was a Phd student in Math/ISyE department. At the same time I am a Business Analytics Manager at a corporation.

I've finished all requirements for graduating except for thesis before putting it on hold and start working.

Right now, my plan is to continue my job and resume research at the same time. My problem is that I have been away from research for 2.5 years and the research topic that I did is no longer interesting.

I'd like to pick your brain on the following question:
what research topic is going to help me in my career after I achieve my PhD (if I ever finish it). I was thinking of a few, model/tool/technique on Revenue Management, Forecasting, Multivariate involving big data, optimization? Techniques that involve banking models (they get high salary at banks, right)?
Honestly I'm not sure what will be "hot" in business, that will help me in the corporation world.

A little bit about my background: I have quite strong math background, and used to do research in non-parametric statistics, reliability theory.

Thank you for any of your contributions!
Hey StatPhD.

This is just my own opinion (so take it for what its worth), but I imagine that with an increase in computing power, corporations will be more and more interested in knowing how to find patterns in both structured and unstructured data environments.

Anything that results in an increase in this area (in terms of its effectiveness in finding patterns) will be something that corporations will value in some form or another.

This is of course a very broad area, and what you will have to do is work on a particular domain and data set.

Also be aware that if the executives and decision makers don't like using anything advanced, then all the great new research and techniques won't mean a thing to them: if you are working in the banking environment or the tech environment (like hedge funds and IB), then this shouldn't be a problem.

Another thing that is equally important is knowing when the data is not representative of a particular attribute and when it is highly biased. This is crucial when you are taking public data and trying to make sense of it (which I'm sure many banks and data mining organizations are interested in).

Hopefully these give you some things to think about.