0

R is good at processing data. It is an analytic turbo-ginsu.

So LabVIEW is great at getting data. It plugs into anything with electrodes.

MatLab has simulink. There are toolboxes for data acquisition.

What does "R" have? Does it connect to hardware, or is that too "laboratory" for it? What is the "R" interface to hardware for gathering the data that it so beautifully analyzes?

There is ephemeris associated with the method of collection and with the method of analysis that can be mutually informative. If I have video input, and I want to do 2d image tracking, why would I use R? I wouldn't because it can't inform the end-product. It has great libraries that traverse the breadth of the field of mathematics, that are made by the best in the field so they tend to be substantially correct. That doesn't mean anything in a product-centric world. You have to be able to productize your work. MATLAB has the MATLAB compiler. LabVIEW is compiled and has standalone runtimes. Both connect to hardware on one side and the user on the other with good math in the middle. R doesn't have that - or does it?

Package list, documentation, and examples are requested for this answer.

EngrStudent
  • 8,232
  • 2
  • 29
  • 82
  • It is about integration with the data source. Philosophically speaking - there is value in "vertical integration" between the analysis tool and the data source. Things like precision, sampling rates, and tool ephemeris are implicit in that connection and can be informative in an analysis. Relevant keywords include: "collecting data" and "data-driven computing". – EngrStudent Jul 11 '14 at 15:43
  • This question is about the facilities that R offers, not about statistics per say (even though those facilities might be part of the process that leads to statistical analyses). Although R is my favorite statistical software, we are not here to 'sell' you on using it for your products. This question is off-topic for CV. – gung - Reinstate Monica Jul 11 '14 at 18:55
  • R is wrapped in others products, like SQL-pathfinder. Swallowed whole. Also it is in JMP. It is to some degree wrappered by Python. These are wholescale productizations of it. Should I delete this question? – EngrStudent Jul 11 '14 at 18:57
  • You don't have to delete the question, if you don't want to. (Actually, once there's an answer w/ a positive score, I'm not sure if SE will let you, dk.) You are right that there are wrappers for R in commercial products, but the Q isn't on-topic here (which doesn't make it a bad Q). You could try on something like 1 of the r-help listservs. – gung - Reinstate Monica Jul 11 '14 at 19:10
  • The reason this is off topic is because answers will be opinion based. For example - in my opinion - it's best to use labview for data acquisition, compiled matlab for pseudo real time analysis, and R for data exploration. This kind of works for me in most of my projects. Or at least I make it work. Anyway, that's my opinion: there is not a definitive answer to your question as currently posted. – Andy Clifton Jul 13 '14 at 19:22
  • This isn't asking for opinion. It is asking for a single example of using "R" only both for data acquisition (not simulation) and data analysis. – EngrStudent Jul 14 '14 at 11:27

1 Answers1

2

You don't need anything but base R. Use the scan( ) function and pipe from LabVIEW to stdio.

Dennis
  • 1,710
  • 9
  • 15
  • How do I pipe from LabVIEW? Can you elaborate, or give a basic example. – EngrStudent Jul 11 '14 at 09:26
  • 1
    I don't use LabVIEW, so I can't answer your question. In R, look at help(scan) and check your OS syntax for setting up a pipe. In a Linux-like system, you would set LabVIEW to output to what appears to it to be a file, but is really the standard input to R. In R, you use the scan( ) function to read (and parse) the LabVIEW output into a data frame. – Dennis Jul 11 '14 at 14:57