Posted on behalf of one of our members.
Do you have an opinion on SciDB, www.sciDB.org? It appears to be a separate tool that could use R as a wrapper, but where the R packages will not, without difficulty, communicate with the SciDB objects. SciDB uses arrays.
What is the big picture view of Hadoop, MapReduce and SciDB? How do they fit together?
I’m limited on time for education. I don’t need cool tools; I need results.
My 2 cents on focusing on a short doable list of new tools to learn is R, Python, and SciDB. (My baseline toolset now: VBA, SQL and SAS)
My motivation is that the workplace over the next 20 years might require me to get another degree just to stay employed. I only have a master’s in physics, not much background in statistics.
No reason to get a degree with all of the free online courses. Pick some interesting idea, analyze it using RMarkdown and publish it on http://rpubs.com. Poof, instant reference for your expertise. Start blogging about what you learn. Don't worry if 1,000 other people say the same thing. Software constantly changes and adapts. You'll be writing about new stuff overlaying old stuff. If you don't have a specific reason to use SciDB (it's fanatasic, btw), then stick with something more mainstream and approachable like R. You could work your whole life using R and never need another tool.