A Fine Slice Of SQL Server
Cosima Meyer explains how debugging works in R with RStudio:
Three basic commands in RStudio let you do the debugging:
debug(function_name),browser(), andundebug(function_name).With
debug(function_name)you start the debugging of your function – it’s basically like a mole that digs in. When you’re in debug mode, you can also call the objects in your function.
Read the whole thing to learn the power of debugging beyond the print() statement. H/T R-Bloggers.
Sebastian Sauer takes us through an interesting scenario.
No lede here because it’s almost 100% code and headers. A quick description of this is that we can see ways to parse columns in an R DataFrame and plot visuals without hard-coding the column name in our plot definition, using a variable instead.
And I had to rewrite the synopsis above because I used the data science term “variable” until hitting a wall when describing the programming term “variable.”
Comments closedMichael Mayer and Christian Lorentzen team up:
SHAP is one of the most used model interpretation technique in Machine Learning. It decomposes predictions into additive contributions of the features in a fair way. For tree-based methods, the fast TreeSHAP algorithm exists. For general models, one has to resort to computationally expensive Monte-Carlo sampling or the faster Kernel SHAP algorithm. Kernel SHAP uses a regression trick to get the SHAP values of an observation with a comparably small number of calls to the predict function of the model. Still, it is much slower than TreeSHAP.
Read on to see how to do this in both R and Python. With libraries the way they are, the code is very similar and the results are basically the same.
Comments closedThomas Williams waits for the keystroke:
This R Markdown snippet demonstrates “debouncing”: waiting until a user stops changing an input, before updating dependent charts and tables. Debouncing is often used in web sites to prevent the user interface “jumping” as data is being entered, especially when the update takes a noticeable amount of time – for instance calling an API or database, or doing a calculation.
Read on to see an abridged example, as well as a link to the full version.
Comments closedBrendan Tierney notes an update to Oracle Autonomous Database:
Oracle Machine Learning for R (OML4R) is available on Oracle Autonomous Database. Finally. After waiting for way, way too long we can now run R code in the Autonomous Database (in the Cloud). It’s based on using Oracle R Distribution 4.0.5 (which is based on R 4.0.5). This product was previously called Oracle R Enterprise, which I was a fan of many few years ago, so much so I wrote a book about it.
I’m curious how well this has done compared to SQL Server ML Services—my experience is that getting ML Services adoption has been a struggle.
Comments closedThis blog post is a follow on to our API as a package series, which looks to expand on the topic of testing {plumber} API applications within the package structure leveraging {testthat}. As a reminder of the situation, so far we have an R package that defines functions that will be used as endpoints in a {plumber} API application. The API routes defined via {plumber} decorators in
instsimply map the package functions to URLs.
Jamie covers a lot of testing ground in that post as well, so check it out.
Comments closedThomas Williams has a change of font:
Setting a different font is a minor tweak to an R Markdown file that might help it fit better with a corporate or company look. Google Fonts is an industry-accepted method to reference fonts on web pages and can be used with R Markdown rendered by Shiny. There’s no need to download anything while developing the R Markdown file in RStudio, or viewing it in a web browser from a Shiny server.
I’ve recently used this technique to integrate the “Atkinson Hyperlegible” font, a font with “…greater legibility and readability for low vision readers…” from https://brailleinstitute.org/freefont. Since Shiny themes already specify a font, the steps below show how to override that with the Google Font, demo’d with the
spacelabShiny theme.
Click through to see how you can use this font, as well as others, in Shiny apps.
Comments closedThomas Williams wants to improve the user experience:
Recently I needed to reload a parameter, without reloading the page. The parameter was bound to a data frame, where end-users selected a value and then I looked up other fields in the data frame further down the page (for example, a name was selected, but I wanted the identifier from the same record). It wasn’t exactly intuitive, so here’s how I did it.
Click through for code and explanation.
Comments closedJamie Owen continues a series on building a package around an API:
Part 1 of this series laid out some ideas for how one might structure a {plumber} application as an R package, inspired by solutions such as {golem} and {leprechaun} for {shiny}. In this installment of the series we look at adding some functions to our package that will take care of logging as our application runs. If you haven’t already, we recommend reading the first installment of this series as the example package created for that post will form the basis of the starting point for this one.
Read the whole thing.
Comments closedMikolaj Spytek promotes an R package:
You can learn about it in this blog, but long story short, survival models (most often) predict a survival function. It tells us what is the probability of an event not happening until a given time t. The output can also be a single value (e.g., risk score) but these scores are always some aggregates of the survival function and this naturally leads to a loss of information included in the prediction.
The complexity of the output of survival models means that standard explanation methods cannot be applied directly.
Because of this, we (I and the team: Mateusz Krzyziński, Hubert Baniecki, and Przemyslaw Biecek) developed an R package — survex, which provides explanations for survival models. We hope this tool allows for more widespread usage of complex machine learning survival analysis models. Until now, simpler statistical models such as Cox Proportional Hazards were preferred due to their interpretability — vital in areas such as medicine, even though they were frequently outperformed by complex machine learning models.
Read on to dive into the topic. H/T R-Bloggers.
Comments closed