Background: My PhD was in 'Computational Science'. My dissertation was on the analysis of X-Ray Diffraction Data and analysis of thermally perturbed nuclei in the overall dynamic analysis of the molecular electron density for solid state physics. The takeaway? It was very much based in science.

In my opinion Computational Science is the pursuit of science, "... a systematic enterprise that builds and organizes knowledge in the form of testable explanations and predictions about the universe" (wiki), via computational means.

Most position fo 'Data Science', however, seem more like 'data analysis' types of jobs. That is, heavy SQL queries, using pre-built R and Python models (linear regression, etc.) to draw conclusions from structured and unstructured data.

Is Computational Science a superset of Data Science? Are they interchangeable? Is Data Science an actual 'science'? Is Computational Science an actual 'science'?

  • $\begingroup$ I think that the question has some value, but you would need to work it out a little bit. This draft of a report in CSE might be useful. They have some mention about the relation between the two. You might think on a relationship similar to the one between experimental and theoretical science, somehow. $\endgroup$
    – nicoguaro
    Jul 22 '16 at 20:00

They are not interchangeable.

  • Computational science tends to refer more to HPC, simulation techniques (differential equations, molecular dynamics, etc.), and is usually referred to as scientific computing.

  • Data science tends to refer to computationally-intensive data analysis, like "big data", bioinformatics, machine learning (optimization), Bayesian analyses using MCMC, etc. I think it's the same as what used to be referred to as computational statistics. It was the infusion of computer science with statistics, but many of the techniques that were developed dropped the rigorous Fisherian "statistical testing" (clustering, cross-validation techniques, data visualization) but kept the data part.

The most clear explanation of it came to me when I was teaching a workshop on Julia for Data Science and Scientific Computing. The data scientists wanted to learn Julia in order to do fast "big data" analysis, i.e. regressions and other GLMs on large data. The computational scientists (scientific computerers?) wanted to know how to easily write code to solve large linear systems on HPCs and GPUs.

Notice those are two ways of saying the exact same computations, but with very different meanings. So in some sense similar, but still distinct (and there is cross-over between the disciplines, like using machine learning to learn parameters for PDEs from data).


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