[Apologies if this post sounds naive, I'm fairly new to the world of data science/big data and very unsure where I'm heading career-wise]

I'm currently an undergraduate MMath [integrated master's] Mathematics student in the UK who has finished the third year of the course [out of four years].

As I have been considering the possibility of doing further research in Mathematics/Statistics/Operational Research/Data Science, I have decided to stay on and complete the Master's component of the course [as it is the only Master's course I can get funding for at this stage]. After the Master's I may continue on and do a PhD.

There are currently two projects that appeal to me that seem to have relevant applications. The first one is on improved MCMC [Markov Chain Monte Carlo] methods, in particular MCMC using Hamiltonian Dynamics. There is scope for some big data applications here.

The other project that I could take part in is one on the centrality/communities detection of networks within network science. This could possibly be useful with applications in Operational Research.

Does anyone have an idea as to which project will be more relevant to data science/analytics?


1 Answer 1


Both MCMC methods and network analysis play an important role in data science. I think you should go for the project you like more. However, in my experience, community detection in networks is a niche (applied math/graph theory) of network analysis, while MCMC methods involve lots of statistical and computational concepts. Personally, being a statistician, I would go for the MCMC methods.

  • $\begingroup$ My view is that data science is dominated by machine learning and computer scientists and that statisticians take a back seat. One index of this in the States is that the Amer Stats Assoc (ASA) has several blogs on its website about concerns that statistical grant requests are being underfunded, esp relative to CS grants. Not to mention that network analysis is a big thing in data science and obtaining deep, good experience in that would make your instantly employable. $\endgroup$
    – DJohnson
    Commented Jun 16, 2015 at 11:03
  • $\begingroup$ I don't want to start a religion war on this subject, but in my opinion machine learning IS statistics in a broader sense: it is statistical learning. Then, of course a statistician who wants to become a data scientist needs to develop some CS skills. $\endgroup$ Commented Jun 16, 2015 at 11:07
  • $\begingroup$ There you go! You've confirmed my point. But I do agree that opinions can run high on this issue. Again, by pointing to the ASA's own statements, one can get a sense of what is really happening out there and which side is winning the war right now. $\endgroup$
    – DJohnson
    Commented Jun 16, 2015 at 11:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.