Thankyou for this I've completed my bachelor's in stats now thinking of doing my masters, this video really put my love for stats in words. Thank you from India.
I dont understand , how then one can update such programm in a dynmaic or.live setting, i.e. fetching data on a regular basis while embedding the whole thing within ui and server separate shiny app functions? Your input is highly appreciated Merci
@@mikiallen7733 I'd say esquisse is not meant for such settings, e.g. regular data updates. I'd use it to get started writing ggplot2 code. In shiny apps I wouldn't embed esquisse.
@@StatistikinDD no. I think Now a have R version 4.3.2 I would like to update, do I need to Remove old version or I can try it: install.packages ("installr") library(installr) updateR()
@@StatistikinDD after that I Would like to Join two columns become unique column in a dataframe. So the new column might be character and have two categories corresponding to the old two columns names. While the old two columns had number of test observed and another the number of tests expected in the same health facility. These numbers is grouped in week (15 weeks). I want to keep this new column and another new column that have number of tests, and another the week identifications (1,2,3...15) If you have some video about basic r skills with dataframe manipulation, please suggest me
@@albinonhabanga8066 Yes, installr is a good option. However, going over the 4.4 threshold you might need to re-install your packages. No need to remove the old R version.
Hallo wie Interpretiere ich denn die restlichen Effekte dann? D.h. Lernstunden alleine nicht signifikant (0,199 > 0,05) und nur in der Interaktion zwischen Lernstunden und Entspannung kann ich signifikanten Lernerfolg nachweisen?
Im Modell mit Interaktionsterm sind die Einzeleffekte nicht separat zu beurteilen. Ein Teil des Effekts steckt im einzelnen Koeffizienten, ein weiterer Teil im Interaktionsterm. Anschaulicher ist es, konkrete Werte für beide Prädiktoren einzusetzen und Vorhersagen dafür zu berechnen anhand der Regressionsformel.
@@StatistikinDD Danke! Mit p-Wert von 0,199 betrachte und interpretiere ich also auch nicht dessen Regressionskoeffizienten (-0,457), sondern nur den des Interaktionsterms? Gibt es eine Quelle dafür, dass man sich nur Variable des Interaktionsterms ansieht? Wie kann das zustande kommen, dass Lernstunden (x) auf Lernerfolg (y) in diesem Moderationsmodell nicht signifikant ist, wenn angenommen in einem linearen Modell eine Signifikanz zwischen Lernstunden und Lernerfolg (gleicher Datensatz) vorhanden wäre?
@@JessiJ-zz7rp Die Effekte verändern sich, wenn die Moderation mit berücksichtigt wird. Es ist schließlich ein anderes Modell! Alle Kombinationen sind möglich: einzeln signifikant, im Moderationsmodell nicht, oder umgekehrt, oder beide Mal sig. oder nicht sig. Die grafische Interpretation ist oft einfacher als das Grübeln über die (nicht mehr von einander unabhängigen!) Koeffizienten. Das Modell mit Interaktionsterm testet in erster Linie, ob eine (statistisch signifikante) Wechselwirkung vorliegt.
So how: load data into Global Environment. Then proceed as in video. I'm just not sure if I'd like the same data in different packages. But there may be situations where it makes sense. E. g. to avoid dependency on that other package, maybe because it contains a lot more than the data, or because it is difficult to access in your environment.
@@StatistikinDD I previously wrote an article introducing my r package, submitted it to the journal of statistical software, and received the following comment: 2 sample tests are run using TEST_FUN(group1, group2) and no formula version is available (it might be less natural in the multidimensional setting but e.g., Hotelling::hotelling.test implements one which is very clear and easy to use). test_fun is mentioned here, but I am not clearly how to use it
Da ist was falsch! Wir "sortieren die Vermögen aller in aufsteigender Reihenfolge" und stellen fest, dass die untere Hälfte weniger zur Verfügung hat? Für den Median spielt das keine Rolle, aber die Aussage ist nicht richtig...
i'm learning sql but want to do everything in R for convenience. this was very informative as i've seen many other videos show different approaches and packages
Some Excel files are a real challenge to deal with. I found the unheadr package to be helpful in combination with some of the tidyverse packages (tidyr). It also helps to know a little bit about regular expressions (regex).
Thank you so much. I knew about Quarto when it came out, but I needed convincing to learn it. I read some posts about Quarto to see if I need to use it, but I'm completely convinced after watching your video. Your video is short, but it covers all valuable points.
Have you specified a dataset? Is the data suitable? Sometimes there may be conflicts with other packages. In such cases, closing Rcmdr, starting a new R session and reopening Rcmdr worked for me.
Thanks, Joachim! On the one hand, it's fun to dive into new technology. On the other hand, it was another case of "Well, I guess I can throw away my training material on Reporting" when Quarto appeared on the scene ... 🙂
Sie erinnern mich an meinen Lieblingsdozenten aus der Uni! Ich habe alle Funktionen schon in GitHub muss jetzt irgendwie richtig verstehen, wie ich diese in das Projekt bringe
Hi this is really helpful, I was able to follow everything but when documenting the functions I got an error on "sew" in package knir . the man file wasn't created. I have 3 functions in one script
Hört sich nach Apple an. Habe keine Erfahrung mit Rcmdr auf iOS. Vielleicht helfen die Hinweise hier: www.john-fox.ca/RCommander/installation-notes.html
Thank you so much for loading the videos, they have been super helpful! I tried Build -> More -> Load All and got this error message. The folder did have the processx and processx.rdb files, so I'm not sure what the error means. Could you please advise? Error in gzfile(file, "rb") : cannot open the connection In addition: Warning message: In gzfile(file, "rb") : cannot open compressed file '\\mypackages/processx/R/processx.rdx', probable reason 'No such file or directory' Error: unable to load R code in package ‘processx’meaasge
Thanks, these videos are very good and I have created a package by following your instructions. My package is in a private github repo now. May I ask how do other users now access this library and use the functions? Do you have a video for that? Thank you very much!
To install from a private repo, you (i. e. potential users of your package) need to provide authentication. I wouldn't spread authentication information, so if I wanted others to access my package, I'd make the repo public. Maybe it helps if you have a private repo where you try out things, and a public repo containing a version you feel ok about being public?
Hi, guys. I got a strange problem. Some shortcuts work in Rmarkdown, but not in Quarto. For example, Insert Assignment Operator, Insert Pipe Operator. Anyone runs into the same problem?
Hi, haven't noticed that. What is your setup? Operating system, R version, quarto version, RStudio (?) version. For the quarto version, you can install the quarto R package. (Not a replacement for the quarto software, just a small interface.) Then use quarto::quarto_version().
@@StatistikinDD Thanks a lot. Problem solved by updating Quarto. Thanks man. But one more question: is there a way to automatically update the Quarto? I don't even know my Quarto is out of date if you did not mention it.
@@CanDoSo_org Not aware of an automatic update, not from within R anyway. Quarto gets bundled with RStudio, so when you update that, you may get a new quarto version as well. You can install daily builds of RStudio to get quicker updates. (I don't do that.) Apart from that, I sometimes check the quarto website and manually download quarto.
Thanks this is great! Would you know what the problem is with Error in .get_package_metadata(pkgdir) : Files 'DESCRIPTION' and 'DESCRIPTION.in' are missing
@@StatistikinDD Thanks for your prompt reply! yes i do have a DESCRIPTION text file at the top level and I am creating the package using RStudio - New Project - Package.
@@user-mh2pu1pe2n Looks like a known issue. github.com/rstudio/rstudio/issues/12945 github.com/r-lib/devtools/issues/2512 Maybe try updating as much as possible, e. g. RStudio, devtools, pkgbuild
Thank you very much for this video! I am at the starting stage of my career and have only done data analysis from within a large corporation. However, there were many aspects of your experience that I could relate to like inheriting code and having to deal with Office format deliverables. I will follow you on twitter.
Good question. The DV is hard-coded in the function and so not passed on as a function parameter. Line 14: model <- as.formula(paste("result ~", IV)) DV = result for all the models calculated here. Only the IV varies.
@@anmolpardeshi3138 Functions can access R's Global Environment. The result variable (DV) is part of the dataset which is specifiied in the call to lm(), which accesses Global Environment. Note that, however, the workers for running the loops in parallel do not automatically have access to the Global Environment in this approach, at least on Windows. Each worker starts in an empty environment. So the data (and the reg function) are exported to the workers using the clusterExport function in this line of code: clusterExport(cl, c("reg", "data")) cl is the object resulting from the makeCluster call. I'd now recommend this approach: cl <- makeCluster(parallelly::availableCores(omit = 1)) See comment below and a later video on why to avoid detectCores(). ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-EYpuUKsGqPQ.html
Thanks for the video. very helpful. Looks promising, but for use cases where controlled page layout is critical (posters, "one pagers", brochures, etc), it seems maybe we'll have to wait for paged HTML support ala pagedown?
Currently I'm lacking experience in that. Read that fixed width and toggling with self-contained / embed-resources may help in PDF output stability. Thanks to Nicola Rennie.
Thanks for this brillian tutotial, I have a quesstion though: is it ok that I save the generated .tar file in my R package directory on my Desktop? Before ubmitting to CRAN. Because I assume that if I save the .tar file elsewhere, and submit my directory to CRAN, the vignette won't come up on the help tab, right? THANKS!
Hi, thanks for your feedback! For CRAN submission, you upload a tar.gz file to CRAN, which contains the vignette. From the R Packages book at r-pkgs.org/vignettes.html#sec-vignettes-how-built-checked Prepare built vignettes for a CRAN submission: Don’t try to do this by hand or in advance. Allow vignette (re-)building to happen as part of devtools::submit_cran() or devtools::release(), both of which build the package.
Why do you waste so much time on explaining what the function inhouse_calc does? No one cares it scales, no one cares it's not exactly between 0 and 100, etc. The video title says this is about including data in your R package so it would be better to focus on that and leave everything distracting aside.
You have a very logical point but some information about other stuff comes with main topic's wind. You don't just go from point a to b. You also see where the supermarket on your way to point b. This kind of informations is crucial for some people, and showing it with a realistic and alive example is also important. You can simply skip it. Anyways. Great tutorial @StatistikinDD keep it rooling!
Thank you very much for your videos, I subscribed to the channel! Lately, I've been trying to convince my friends and colleagues in a subliminal way to go from Rmarkdown to Quarto. The arguments you presented are exactly what I was missing, thank you! I also have 2 questions: 1. Xarigan presentations are what prevent me from transferring all my remaining rmd to qmd. You said that reveal.js works pretty well on Quarto, is there a feature that is missing compared to Xarigan? 2. You said that Quarto could support languages that do not yet exist yet. I am also a Nim user, is it possible to use Nim in it and what would be the missing functionalities since the language has not been implemented yet? Again thank you for your video!
Thank you for your comment! Much appreciated. I still use xaringan when I need a pdf version. Had difficulties printing quarto/ reveal.js to pdf. Heard fixed width and toggling self-contained true/false could help. Haven't tried yet. For HTML, I really like reveal.js. Can't say about Nim, sorry. Maybe ask Posit forum.
Where I can find a library in r which does very good visuals for the results of quantile regression ? More specifically for plotting which factors affect which quantile / sub group out of the whole sample ?