There is no equivalent to shiny or rstudio server in python.Jupyter
Not the same, have you used them at all?
Why doesn't R ever get any love?
-
I write my code in R and use a small sample of my data so speed doesn’t really matter, just getting the programming correct. Then I find bottle necks and rewrite using rcpp. I’m not very good at that but I can at least make sure I’m getting the same result. Then I run full data and it’s fast enough for everything I need
-
It’s always an exclusive OR in these threads. What prevents from using R, Stata, Julia, Matlab, C++ depending on your problem?
Yes in practice nearly everyone uses several languages. The above was mainly examples for those who boldly claim that R is fast enough. It certainly is for running regressions and related statistical models which leverage native linear algebra routines. But if you implement your own models, especially the kind that need to be approximated with Monte Carlo techniques, soon R becomes much too slow.
-
It’s always an exclusive OR in these threads. What prevents from using R, Stata, Julia, Matlab, C++ depending on your problem?
Yes in practice nearly everyone uses several languages. The above was mainly examples for those who boldly claim that R is fast enough. It certainly is for running regressions and related statistical models which leverage native linear algebra routines. But if you implement your own models, especially the kind that need to be approximated with Monte Carlo techniques, soon R becomes much too slow.
And to be clear, this is a problem shared by all scripting environments and is not specific to R. Julia seems to strike a good balance of "scripting-like" environment and performance; however in my limited experience with some toy models I found that one essentially has to write C-like code to get C-like performance, and that removes a lot of the attractiveness of the platform. The ecosystem is also not yet mature; just think about the mountains of proven, native code with zero overhead available for C and Fortran (LAPACK, GNU Scientific Library, etc).
-
There is no equivalent to shiny or rstudio server in python.
Jupyter
Not the same, have you used them at all?
Yes, not Jupyter, there are Python equivalents to RStudio however, in fact RStudio claims to be one of them. The Python equivalents to R Shiny would be Dash and Streamlit. so there are Python equivalents to these things. Jupyter is equivalent roughly to R Notebooks.
-
At a certain point ggplot is ms paint- I’m feeding it a list of values hoping it renders. At that point, excel is better
or t(matrix(1:9, 3, 3)). Also what kind of projects are you working on that inputting a small matrix is the bottleneck? I don't think anything matches dplyr and ggplot2 in ease of use for data manipulation and plotting. >
If you want your plots to look like Excel plots, Excel is better
Actually there is an Excel theme available for ggplot.
-
At a certain point ggplot is ms paint- I’m feeding it a list of values hoping it renders. At that point, excel is better
or t(matrix(1:9, 3, 3)). Also what kind of projects are you working on that inputting a small matrix is the bottleneck? I don't think anything matches dplyr and ggplot2 in ease of use for data manipulation and plotting. >
If you want your plots to look like Excel plots, Excel is better
Actually there is an Excel theme available for ggplot.
Lol. Of course there is.
-
Bros, get a grip. We are economists, not mathematicians nor computer scientists. We wouldn’t know programming if ot was banging our inflatable gf in the middle of the kitchen table while we were eating our spaghettios. Knock off the nonsense, you are embarrassing us
-
Bros, get a grip. We are economists, not mathematicians nor computer scientists. We wouldn’t know programming if ot was banging our inflatable gf in the middle of the kitchen table while we were eating our spaghettios. Knock off the nonsense, you are embarrassing us
Some of us are computational economists. which is basically economics + scientific programming. It's not your production ready commercial code but then it doesn't have to be.
-
yup. R is very good for this type of work. Obviouly not great for building software like python.
I write my code in R and use a small sample of my data so speed doesn’t really matter, just getting the programming correct. Then I find bottle necks and rewrite using rcpp. I’m not very good at that but I can at least make sure I’m getting the same result. Then I run full data and it’s fast enough for everything I need