Re: [R] dput sparseMatrix list
> On Jan 4, 2016, at 11:17 PM, Lietz, Haiko wrote:
>
> hi all,
>
> when dputting a list of sparse matrices (Matrix package), the output does not
> contain the data but the information that the list contains sparse matrices.
>
> M <- sparseMatrix(i = c(2, 1), j = c(1, 2), x = c(1, 1))
>
> dput(M) ... works.
>
> dput(list(M, M)) ... does not work.
>
> how can I dput a list of sparse matrices?
> MM <- list(M,M)
> dput(MM)
list(,
)
No problem.
I do get an error (clarifying the "did not work" statement), as (perhaps) did
you?
> dput(M,M)
Error in cat("new(\"", clx, "\"\n", file = file, sep = "") :
invalid connection
Perhaps the `dput` function was not configured to handle two S4 objects in a
list?
> [[alternative HTML version deleted]]
Please do better next time.
>
> __
> R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
David Winsemius
Alameda, CA, USA
__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
[R] dput sparseMatrix list
hi all, when dputting a list of sparse matrices (Matrix package), the output does not contain the data but the information that the list contains sparse matrices. M <- sparseMatrix(i = c(2, 1), j = c(1, 2), x = c(1, 1)) dput(M) ... works. dput(list(M, M)) ... does not work. how can I dput a list of sparse matrices? thx haiko [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] rZeppelin: Easy Spark for R Data Scientists
rZeppelin is an R Interpreter for the Apache (Incubating) Zeppelin project. The intention of rZeppelin is to make it possible for regular R-using non-programmer to integrate the power of Spark, and the wide range of ML packages available for Python and scala, into their day-to-day toolbox — without having to learn a new language, without any learning curve beyond a review of the SparkR API, and without the budget needs or administrative overhead of setting up a Spark or hadoop infrastructure. Zeppelin is a notebook (like iPython) built on top of Spark. Zeppelin provides interactive data visualization and other features, and interpreters for a wide variety of “big data” stores. rZeppelin makes it possible to combine R, scala, and Python code in a single data/ML pipeline, seamlessly, from a single, familiar, interface. (And without breaking lazy evaluation!) This means that you can use the Spark package-base of ultra-fast implementations of popular ML algorithms optimized for clusters, as well as python packages, as an extension of your existing work with R. For example, imagine loading text data in R, running LDA on the text using the distributed implementation of LDA in Spark’s MLLIB, tagging the text using advanced Python NLP packages such as gensim, and then visualizing and further processing the results in R — all from the same interface, in the same session. rZeppelin lets you do this because the R interpreter, along with Zeppelin’s scala and Python interpreters, share the same Spark backend. Apart from Spark, most common datatypes can be moved among R, scala, and Python through the “ZeppelinContext,” a shared environment. rZeppelin is integrated with Zeppelin’s interactive visualization features. It also uses knitr for compatibility with most R data visualization and interactive visualization packages, such as ggplot2 and rCharts. rZeppelin is available here: https://github.com/elbamos/Zeppelin-With-R [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] R package built using newer version of R
On 04/01/2016 2:02 PM, Tyler Auerbeck wrote:
We're currently looking at using the R eclipse plugin StatET as our
development environment. Due to certain requirements, we're still using
2.15.1. However a required package of StatET was built using 2.15.3, which
results in the following warning:
Warning message:
package 'rj' was built under R version 2.15.3
I'm still fairly new to R, but is there any way for us to rebuild this
package using 2.15.1? It doesn't appear to cause us any issues, but it's
still not desirable for users to see that warning.
Any help would be appreciated.
Yes, it's quite easy to do so. StatET probably gives menu options to do
it, but I don't know them: you might want to ask them. From the R
console, try
install.packages("pkgname", type="source")
and if you have the necessary prerequisites (e.g. compilers), you'll get
it installed from source. If it fails, post the errors and the results
of sessionInfo() here, and we'll probably be able to tell you what to do
next.
Duncan Murdoch
__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
[R] R package built using newer version of R
We're currently looking at using the R eclipse plugin StatET as our development environment. Due to certain requirements, we're still using 2.15.1. However a required package of StatET was built using 2.15.3, which results in the following warning: Warning message: package 'rj' was built under R version 2.15.3 I'm still fairly new to R, but is there any way for us to rebuild this package using 2.15.1? It doesn't appear to cause us any issues, but it's still not desirable for users to see that warning. Any help would be appreciated. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] problem in installing package xpose4 in Rstudio
> On Jan 3, 2016, at 11:19 PM, swati j wrote:
>
> With R, package xpose4 is working well, but when I open Rstudio and try to
> install package xpose4
>
> following error message is displayed
>
>> install.packages("C:/Users/om/Downloads/xpose4_4.5.3.tar.gz", repos =
>> NULL, type = "source")
> Installing package(s) into ‘C:/Users/om/Documents/R/win-library/2.15’
> (as ‘lib’ is unspecified)
> ERROR: dependencies 'gam', 'Hmisc' are not available for package 'xpose4'
^ ^ ^ ^ ^ ^
| | | | ^ | |
Please READ error messages. Don't just freak out when you see the work "error".
The rest of the message has meaning.
--
David.
> * removing 'C:/Users/om/Documents/R/win-library/2.15/xpose4'
> Warning in install.packages :
> running command 'C:/PROGRA~1/R/R-215~1.1/bin/i386/R CMD INSTALL -l
> "C:/Users/om/Documents/R/win-library/2.15"
> "C:/Users/om/Downloads/xpose4_4.5.3.tar.gz"' had status 1
> Warning in install.packages :
> installation of package ‘C:/Users/om/Downloads/xpose4_4.5.3.tar.gz’ had
> non-zero exit status
>
> please help me to sort out this problem.
>
> Swati Jaiswal
>
> PhD Scholar (CSIR-Senior Research Fellow)
>
> Pharmacokinetics & Metabolism Division
>
> CSIR-Central Drug Research Institute
>
> *Lucknow-226031, India*
>
> *Mobile +91 9473837970*
>
> [[alternative HTML version deleted]]
>
> __
> R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
David Winsemius
Alameda, CA, USA
__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
[R] problem in installing package xpose4 in Rstudio
With R, package xpose4 is working well, but when I open Rstudio and try to
install package xpose4
following error message is displayed
> install.packages("C:/Users/om/Downloads/xpose4_4.5.3.tar.gz", repos =
> NULL, type = "source")
Installing package(s) into ‘C:/Users/om/Documents/R/win-library/2.15’
(as ‘lib’ is unspecified)
ERROR: dependencies 'gam', 'Hmisc' are not available for package 'xpose4'
* removing 'C:/Users/om/Documents/R/win-library/2.15/xpose4'
Warning in install.packages :
running command 'C:/PROGRA~1/R/R-215~1.1/bin/i386/R CMD INSTALL -l
"C:/Users/om/Documents/R/win-library/2.15"
"C:/Users/om/Downloads/xpose4_4.5.3.tar.gz"' had status 1
Warning in install.packages :
installation of package ‘C:/Users/om/Downloads/xpose4_4.5.3.tar.gz’ had
non-zero exit status
please help me to sort out this problem.
Swati Jaiswal
PhD Scholar (CSIR-Senior Research Fellow)
Pharmacokinetics & Metabolism Division
CSIR-Central Drug Research Institute
*Lucknow-226031, India*
*Mobile +91 9473837970*
[[alternative HTML version deleted]]
__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Re: [R] Better scrolling feature in ggplot using Shiny???
Server-side rendering of large amounts of data is often criticized this way. In general, the answer lies in client-side rendering, which these days usually means serving a Web page with embedded data and Javascript (e.g. D3), not ggplot images. The drawback seems to be a significant amount of extra effort and learning JS (off-topic here), or paying someone else to do that grunt-work. The CRAN Task View on graphics is a little dated in this respect, but may have some options if Web pages are not required. The "plotly" package works with the plotly Web service, but your tool then becomes tied with that service and its licensing requirements, though they do make it easier to get an interactive plot. The "googleVis" package offers some similar features, with similar baggage. Please (re-)read the Posting Guide mentioned at the bottom of every r-help mailing list, which for one thing mentions that this is a plain text mailing list. Posting in HTML is bound to lead to corrupted communication (us not being able to decipher your post) sooner or later, and only you can prevent that by adjusting your email client when you send to this list. [1] https://cran.r-project.org/web/views/Graphics.html -- Sent from my phone. Please excuse my brevity. On January 4, 2016 2:47:19 AM PST, Kunal Shah wrote: >Hello, > >I have plotted a ggplot of large data around 3 points. I opened it >in >Shiny. I want a scrolling feature so that I can just scroll the data. > >I tried to write a code in Shiny where the user can select the slider >range. But "scrolling" by that is not efficient and not at all smooth > >Any help is appreciated > > >Regards > > [[alternative HTML version deleted]] > >__ >R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see >https://stat.ethz.ch/mailman/listinfo/r-help >PLEASE do read the posting guide >http://www.R-project.org/posting-guide.html >and provide commented, minimal, self-contained, reproducible code. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Estimating MA parameters through arima or through package "dlm"
Hi: I don't have time to look at the details of what you're doing but the
"equivalence"
between state space and arima ( as paul gilbert pointed out a few weeks ago
) is not a true equivalence.
if you are in an area of the parameter space that the state space
formulation
can't reach, then you won't get the same parameter estimates. so, what
you're doing
might be okay or might not be, depending on whether the state space
formulation
can reach that area of the parameter space. there's another state space
formulation that is truly equivalent which is called the SSOE formulation
or innovations representation but
I don't know if you want to get into that. google "SSOE state space" if
you're interested.
Mark
On Mon, Jan 4, 2016 at 9:25 AM, Stefano Sofia <
stefano.so...@regione.marche.it> wrote:
> Dear list users,
> I want to use apply a MA(2) process (x=beta1*epsilon_(t-1) +
> beta2*epsilon_(t-1) + epsilon_(t)) to a given time series (x), and I want
> to estimate the two parameters beta1, beta2 and the variance of the random
> variable epsilon_(t).
>
> If I use
> MA2_1 <- Arima(x, order=c(0,0,2))
> I get the following result
>
> [1] "MA2_1"
> Series: x
> ARIMA(0,0,2) with non-zero mean
>
> Coefficients:
> ma1 ma2 intercept
> -0.0279 0.0783 5.3737
> s.e. 0.0667 0.0622 0.0245
>
> sigma^2 estimated as 0.1284: log likelihood=-92.63
> AIC=193.25 AICc=193.43 BIC=207.11
> [1] 0 2 0 0 1 0 0
>
> From this straightforward analysis V[epsilon]=0.1284, beta1=-0.0279 and
> beta2=0.0783.
>
> I also tried to use a DLM representation of ARIMA models and estimate the
> unknown parameters by maximum likelihood through the dlm package (in
> particular applying the example at section 3.2.6, page 115, of "Dynamic
> Linear Models with R" by Petris, Petrone and Campagnoli:
>
> arma_parameters <- function(x)
> {
> buildGap <- function(u)
> {
> gap <- dlmModARMA(ma = u[2 : 3], sigma2 = u[1])
> return(gap)
>}
>init <- c(0.005, 0.004, 0.003)
>outMLE <- dlmMLE(x, init, buildGap)
>dlmGap <- buildGap(outMLE$par)
> }
>
> and this gives:
> [1] "outMLE"
> $par
> [1] 1.00816794 0.02349296 0.02364788
>
> $value
> [1] 3089.196
>
> $counts
> function gradient
> 10 10
>
> $convergence
> [1] 0
>
> $message
> [1] "CONVERGENCE: REL_REDUCTION_OF_F <= FACTR*EPSMCH"
>
> [1] "dlmGap"
> $FF
> [,1] [,2] [,3]
> [1,]100
>
> $V
> [,1]
> [1,]0
>
> $GG
> [,1] [,2] [,3]
> [1,]010
> [2,]001
> [3,]000
>
> $W
>[,1] [,2] [,3]
> [1,] 1.00816794 0.0236848488 0.0238410337
> [2,] 0.02368485 0.0005564272 0.0005600964
> [3,] 0.02384103 0.0005600964 0.0005637899
>
> $m0
> [1] 0 0 0
>
> $C0
> [,1] [,2] [,3]
> [1,] 1e+07 0e+00 0e+00
> [2,] 0e+00 1e+07 0e+00
> [3,] 0e+00 0e+00 1e+07
>
> In this case
> V[epsilon]=W[1,1]=1.00816794
> beta1=W[2,1]/W[1,1]=0.02349296
> beta2=W[3,1]/W[1,1]=0.02364788
>
> I presume that these two approaches should give comparable results, but
> this does not happen.
> Is the model that I used correct? And does it make sense to perform this
> kind of comparison?
>
> This is the log of a rainfall time series (which has already been
> deseasonalised):
> [1] 6.014937 4.978801 5.654592 5.616771 5.612398 5.837147 5.121580 5.832176
> [9] 5.205654 5.355642 5.405376 6.257859 5.516247 5.500850 4.708629 5.482304
> [17] 5.689684 5.727824 4.779123 5.289277 5.217107 5.976351 4.630838
> 5.683240
> [25] 5.345678 5.906179 5.605434 5.497578 5.898801 5.660875 5.111988
> 5.571013
> [33] 5.949340 5.374352 4.841033 5.995706 5.661223 5.458734 4.454347
> 5.795754
> [41] 5.995706 5.596939 5.399971 5.908898 5.282696 5.438514 5.528635
> 6.022721
> [49] 5.524257 5.519459 4.957235 5.547518 5.080783 5.411200 5.056883
> 5.798183
> [57] 5.086361 5.536547 5.220356 5.141664 5.847017 5.052417 5.734635
> 5.340419
> [65] 5.724238 5.634432 5.685958 5.307773 5.817706 5.134032 4.987708
> 5.110179
> [73] 5.423628 5.347108 4.859037 5.556828 5.487283 5.661223 5.732370
> 5.469325
> [81] 5.726848 5.419207 5.172187 5.608006 5.130490 5.586874 5.171052
> 5.683240
> [89] 4.674696 5.286245 5.342813 5.370638 5.432411 5.748118 6.355239
> 5.557986
> [97] 5.399067 5.222516 5.279644 5.425390 5.540871 5.917818 5.132853
> 5.689007
> [105] 5.900993 5.007296 5.102911 5.778271 5.318120 5.927726 5.066385
> 5.716699
> [113] 5.511815 4.714921 5.383577 5.319100 5.269403 5.354698 5.145749
> 5.204556
> [121] 5.878296 5.070161 5.441552 5.213304 5.450180 5.695750 4.893352
> 5.425390
> [129] 5.682559 5.487283 4.213608 5.751620 5.432411 5.379436 5.700444
> 5.580484
> [137] 5.357529 5.319100 4.532599 5.603225 5.208393 5.254888 5.017280
> 5.349961
> [145] 4.374498 5.187944 5.585374 5.716370 3.561046 5.119789 5.163070
> 5.422745
> [153] 5.863915 5.651436 4.762174 5.655642 4.797442 5.735927 4.911183
> 5.240688
> [161] 5.148076 5.477300 4.572647 5.493473 5.437644 4.854371 4.908233
> 4.755313
> [169] 5.582744 5.527841 5.613128 5.211124 5.
[R] Estimating MA parameters through arima or through package "dlm"
Dear list users,
I want to use apply a MA(2) process (x=beta1*epsilon_(t-1) +
beta2*epsilon_(t-1) + epsilon_(t)) to a given time series (x), and I want to
estimate the two parameters beta1, beta2 and the variance of the random
variable epsilon_(t).
If I use
MA2_1 <- Arima(x, order=c(0,0,2))
I get the following result
[1] "MA2_1"
Series: x
ARIMA(0,0,2) with non-zero mean
Coefficients:
ma1 ma2 intercept
-0.0279 0.0783 5.3737
s.e. 0.0667 0.0622 0.0245
sigma^2 estimated as 0.1284: log likelihood=-92.63
AIC=193.25 AICc=193.43 BIC=207.11
[1] 0 2 0 0 1 0 0
From this straightforward analysis V[epsilon]=0.1284, beta1=-0.0279 and
beta2=0.0783.
I also tried to use a DLM representation of ARIMA models and estimate the
unknown parameters by maximum likelihood through the dlm package (in particular
applying the example at section 3.2.6, page 115, of "Dynamic Linear Models with
R" by Petris, Petrone and Campagnoli:
arma_parameters <- function(x)
{
buildGap <- function(u)
{
gap <- dlmModARMA(ma = u[2 : 3], sigma2 = u[1])
return(gap)
}
init <- c(0.005, 0.004, 0.003)
outMLE <- dlmMLE(x, init, buildGap)
dlmGap <- buildGap(outMLE$par)
}
and this gives:
[1] "outMLE"
$par
[1] 1.00816794 0.02349296 0.02364788
$value
[1] 3089.196
$counts
function gradient
10 10
$convergence
[1] 0
$message
[1] "CONVERGENCE: REL_REDUCTION_OF_F <= FACTR*EPSMCH"
[1] "dlmGap"
$FF
[,1] [,2] [,3]
[1,]100
$V
[,1]
[1,]0
$GG
[,1] [,2] [,3]
[1,]010
[2,]001
[3,]000
$W
[,1] [,2] [,3]
[1,] 1.00816794 0.0236848488 0.0238410337
[2,] 0.02368485 0.0005564272 0.0005600964
[3,] 0.02384103 0.0005600964 0.0005637899
$m0
[1] 0 0 0
$C0
[,1] [,2] [,3]
[1,] 1e+07 0e+00 0e+00
[2,] 0e+00 1e+07 0e+00
[3,] 0e+00 0e+00 1e+07
In this case
V[epsilon]=W[1,1]=1.00816794
beta1=W[2,1]/W[1,1]=0.02349296
beta2=W[3,1]/W[1,1]=0.02364788
I presume that these two approaches should give comparable results, but this
does not happen.
Is the model that I used correct? And does it make sense to perform this kind
of comparison?
This is the log of a rainfall time series (which has already been
deseasonalised):
[1] 6.014937 4.978801 5.654592 5.616771 5.612398 5.837147 5.121580 5.832176
[9] 5.205654 5.355642 5.405376 6.257859 5.516247 5.500850 4.708629 5.482304
[17] 5.689684 5.727824 4.779123 5.289277 5.217107 5.976351 4.630838 5.683240
[25] 5.345678 5.906179 5.605434 5.497578 5.898801 5.660875 5.111988 5.571013
[33] 5.949340 5.374352 4.841033 5.995706 5.661223 5.458734 4.454347 5.795754
[41] 5.995706 5.596939 5.399971 5.908898 5.282696 5.438514 5.528635 6.022721
[49] 5.524257 5.519459 4.957235 5.547518 5.080783 5.411200 5.056883 5.798183
[57] 5.086361 5.536547 5.220356 5.141664 5.847017 5.052417 5.734635 5.340419
[65] 5.724238 5.634432 5.685958 5.307773 5.817706 5.134032 4.987708 5.110179
[73] 5.423628 5.347108 4.859037 5.556828 5.487283 5.661223 5.732370 5.469325
[81] 5.726848 5.419207 5.172187 5.608006 5.130490 5.586874 5.171052 5.683240
[89] 4.674696 5.286245 5.342813 5.370638 5.432411 5.748118 6.355239 5.557986
[97] 5.399067 5.222516 5.279644 5.425390 5.540871 5.917818 5.132853 5.689007
[105] 5.900993 5.007296 5.102911 5.778271 5.318120 5.927726 5.066385 5.716699
[113] 5.511815 4.714921 5.383577 5.319100 5.269403 5.354698 5.145749 5.204556
[121] 5.878296 5.070161 5.441552 5.213304 5.450180 5.695750 4.893352 5.425390
[129] 5.682559 5.487283 4.213608 5.751620 5.432411 5.379436 5.700444 5.580484
[137] 5.357529 5.319100 4.532599 5.603225 5.208393 5.254888 5.017280 5.349961
[145] 4.374498 5.187944 5.585374 5.716370 3.561046 5.119789 5.163070 5.422745
[153] 5.863915 5.651436 4.762174 5.655642 4.797442 5.735927 4.911183 5.240688
[161] 5.148076 5.477300 4.572647 5.493473 5.437644 4.854371 4.908233 4.755313
[169] 5.582744 5.527841 5.613128 5.211124 5.275049 5.462984 5.016617 5.981919
[177] 5.566817 5.094364 5.314191 5.712742 5.299317 5.452325 4.691348 5.851628
[185] 5.410753 5.488938 5.660179 5.900993 5.380819 5.256453 4.781641 5.531807
[193] 5.497578 5.274537 4.325456 5.271973 5.077047 5.258536 5.280662 5.247024
[201] 5.995208 4.700480 4.991113 5.457029 5.194622 5.487283 5.197391 5.747161
[209] 5.842094 5.372497 5.306781 5.641907 5.565286 5.259057 5.241218 4.759607
[217] 4.550714 5.230574 4.470495 5.664348 4.846547 5.771130 4.823502 5.598422
[225] 5.627621 5.547518 5.596939 5.468482 5.536940 5.606170 5.281680 5.656691
[233] 5.283204 5.752255 5.192401 4.550714
Thank you for your attention and your help
Stefano
AVVISO IMPORTANTE: Questo messaggio di posta elettronica può contenere
informazioni confidenziali, pertanto è destinato solo a persone autorizzate
alla ricezione. I messaggi di posta elettronica per i client di Regione Marche
possono contenere informazioni confidenziali e con privilegi legali. Se non si
è il destinatario specificato, non leggere, copi
[R] Better scrolling feature in ggplot using Shiny???
Hello, I have plotted a ggplot of large data around 3 points. I opened it in Shiny. I want a scrolling feature so that I can just scroll the data. I tried to write a code in Shiny where the user can select the slider range. But "scrolling" by that is not efficient and not at all smooth Any help is appreciated Regards [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
