Some interesting Data Science stuff found between 2018-01-16 and 2018-01-16. https://www.tidyverse.org/articles/2018/01/dbplyr-1-2/ - a new version of the database backend for dplyr. It allows using stringr functions in the mutate statements, and the operations are evaluated directly on the database. #applyrds #db #rstats https://t.co/76sX7KjIxR https://github.com/welovedatascience/stranger - new package for anomaly detection in R. #rstats #pkg #applyrds https://t.co/O1itP9YXML https://hughjonesd.github.io/huxtable/ - an alternative for xtable? I hope so:) Conditional formatting (e.g., make background red if the value is larger than 3) seems to be very easily achievable.

R has various packages to call other languages, like Rccp, rJava or sparklyr. Those tools significantly expand R’s capabilities, because the user doesn’t need to learn a lot of different stuff. Everything is nicely integrated into R. However, sometimes the problem is different - there’s an existing system written in some language, and R can be used to expand its possibilities. So in that scenario R must be called. In that post, I’ll describe how R can be integrated with C# program using Microsoft.

The best way to organize shiny app is to use modules. They are also an excellent choice when you need to replicate some functionality few times. For example, when you want to compare some plots with different parameters, the modules are your way to go. In basic usage, modules require a direct call of callModule in a server and a place in UI. However, sometimes you don’t know how many instances of a given module are needed.

This is the last (planned) RNews in 2017. In this week there’s something about reproducible research (anyone should do this!), and deep learning - one tutorial, and one general article about image classification in radiology. Books http://www.britishecologicalsociety.org/wp-content/uploads/2017/12/guide-to-reproducible-code.pdf - this book is a guide to writing reproducible code using R. There’s an in Ecology and Evolution in the title, but the scope of this writing should not be limited to only that areas:) Articles: https://tensorflow.

This list is quite small because I hit the history limit on Twitter:( Nevertheless, there are some exciting things - e.g. videos from H2O World 2017. Videos https://www.youtube.com/playlist?list=PLNtMya54qvOHQs2ZmV-pPSW_etMUykE0_ - list of all videos from H2O World 2017. There are at least a few videos which seem to be worth watching. Unfortunately, I didn’t have time to do so:(. Articles: https://rviews.rstudio.com/2017/12/11/r-and-tensorflow/ - this article claims that installing keras (for deep learning) is as simple as calling keras::install_keras().

Another pack of news from R’s world. Aggregators: https://rweekly.org/ - this is a great aggregator - even better than mine;) Articles: http://smarterpoland.pl/index.php/2017/12/explain-explain-explain/ - a quick summary of the packages which can be used to explain the results of various models (lm, glm, xgboost, etc.). Unfortunately, there’s nothing about LIME, which is more general purpose package for explaining models (https://cran.r-project.org/web/packages/lime/index.html). https://christophm.github.io/interpretable-ml-book/ - an online book about explaining models predictions.

Next portion of exciting information from R’s world (in fact, not only R’s but it’s the main thing here). Aggregators: https://trello.com/b/rbpEfMld/data-science - a huge collection of the resources related to the Data Science (R, Python, Big Data, and so on). Articles: https://drsimonj.svbtle.com/visualising-residuals - long post about visualization of the residuals. The “Multiple Regression” section is worth to check - the analysis is quite impressive. http://www.tandfonline.com/doi/full/10.1080/01621459.2017.1311264 - the old saga about why p-values are evil continues.

Shiny application is a great way to deliver the result of statistical analysis, especially when it must be reproducible. I don’t know why, but people prefer to use graphic interface, rather than run scripts;) One of my clients recently requested to have an ability to move around elements in the dashboard. There are some R packages to attain such effect, and possibly it would be a bit easier to use them, but I had an unfinished project called dragulaR.

Here is a small pack of some engaging news from R’s world gathered from Twitter in the last week. R Tips http://www.storybench.org/convert-google-doc-rmarkdown-publish-github-pages/ - title speaks for itself - it shows how to convert GoogleDocs to Rmarkdown format. http://blog.jumpingrivers.com/posts/2017/speed_package_installation/ - in short - options(Ncpus = 6) - allows to use multicores in install.packages, which can significantly speed up the packages’ installation process. Packages https://www.tidyverse.org/articles/2017/11/withr-2.1.0/ - allows a user to call code in a special environment with some global variables alerted.

Here is a small pack of some engaging news from R’s world gathered from Twitter in the last week. Articles https://www.joelonsoftware.com/2003/10/08/the-absolute-minimum-every-software-developer-absolutely-positively-must-know-about-unicode-and-character-sets-no-excuses/ - minimal amount of knowledge about Unicode. The article is targeted to developers, but it might also be useful for Data Scientists. I highly recommend other posts on that blog - the host is one of the creators of VBA in Excel, Trello, and StackOverflow. https://imgur.com/gallery/GD5gi - visualization of the sorting algorithms.