Press "Enter" to skip to content

Category: R

Docker And R

Published 2017-11-21 by Kevin Feasel

Mara Averick has some resources to help you get started with running R in a Docker container:

liftr 📦 by Nan Xiao

liftr aims to solve the problem of persistent reproducible reporting. To achieve this goal, it extends the R Markdown metadata format, and uses Docker to containerize and render R Markdown documents.

Click through for those resources as well as an addictive 8-bit animated GIF.

Comments closed

Building Dynamic Row Headers With ML Services

Published 2017-11-17 by Kevin Feasel

Dave Mason tries to get around his RESULT SETS limitation when using SQL Server Machine Learning Services:

The columns in the data frame clearly have names, but SQL Server isn’t using them. The data frame columns have types in R too (more on this in a moment). Now that makes me wonder about the data types for the data returned by SQL. How is that determined? If SQL isn’t using the column names, can I assume it isn’t making use of the R column types either?

For a point of reference, let’s run some more R code to show the column names and types. As before, the rvest package is used to scrape a web page, with each HTML <table> found becoming a data frame in the “tables” list (line 3). A data frame of table metadata is created by calling data.frame(). The first parameter is a vector of column names (line 4), the second parameter is a vector of column classes (line 5), and the third parameter causes the row “names” to be incrementing digits (line 6).

This is a work in progress as Dave continues his series.

Comments closed

Defining Result Sets With ML Services

Published 2017-11-16 by Kevin Feasel

Dave Mason covers a pain point in SQL Server Machine Learning Services:

The example above is so simple, defining the RESULT SETS poses no problems. But what if the format of the output isn’t known at design time? R (or Python) might take the input data set and add, remove, or change columns conditionally. Further, the input data set might not even be known at design time. How would you define the RESULT SETS at run time?

WITH RESULT SETS needs a MAKE_A_GUESS or FIGURE_IT_OUT option. If there’s some other type of “easy button” for this, I haven’t found it.

It would be nice if the service could the ability to read the data frame columns and use those by default.

Comments closed

Interpreting P-Value Histograms

Published 2017-11-15 by Kevin Feasel

David Robinson visualizes and interprets different p-value histograms:

So you’re a scientist or data analyst, and you have a little experience interpreting p-values from statistical tests. But then you come across a case where you have hundreds, thousands, or even millions of p-values. Perhaps you ran a statistical test on each gene in an organism, or on demographics within each of hundreds of counties. You might have heard about the dangers of multiple hypothesis testing before. What’s the first thing you do?

Make a histogram of your p-values. Do this before you perform multiple hypothesis test correction, false discovery rate control, or any other means of interpreting your many p-values. Unfortunately, for some reason, this basic and simple task rarely gets recommended (for instance, the Wikipedia page on the multiple comparisons problem never once mentions this approach). This graph lets you get an immediate sense of how your test behaved across all your hypotheses, and immediately diagnose some potential problems. Here, I’ll walk you through a basic example of interpreting a p-value histogram.

It’s a fun read and informative as well.

Comments closed

A Compendium Of R Errors

Published 2017-11-15 by Kevin Feasel

Sumendar Karupakala has a bunch of errors you might find in R, as well as their explanations and fixes:

#pull out the animals which are dogs
animaldata[animaldata$Animal.Type == “Dog” ] # throuws an error
Error in `[.data.frame`(animaldata, animaldata$Animal.Type == “Dog”): undefined columns selected
Traceback:
1. animaldata[animaldata$Animal.Type == “Dog”]
2. `[.data.frame`(animaldata, animaldata$Animal.Type == “Dog”)
3. stop(“undefined columns selected”)
In [8]:
#fixed error
animaldata[animaldata$Animal.Type == “Dog”, ] # missedout comma with in the bracket

Some of it is basic syntax; others are a bit nastier.

Comments closed

Curl For Windows Now Uses WinSSL

Published 2017-11-14 by Kevin Feasel

David Smith notes that curl version 3.0 supports the winSSL library rather than installing OpenSSL:

To implement secure communications, the curl package needs to connect with a library that handles the SSL (secure socket layer) encryption. On Linux and Macs, curl has always used the OpenSSL library, which is included on those systems. Windows doesn’t have this library (at least, outside of the Subsystem for Linux), so on Windows the curl package included the OpenSSL library and associated certificate. This raises its own set of issues (see the post linked below for details), so version 3.0 of the package instead uses the built-in winSSL library. This means curl uses the same security architecture as other connected applications on Windows.

This shouldn’t have any impact on your web-connectivity from R now or in the future, except the knowledge that the underlying architecture is more secure. Nonetheless, it’s possible to switch back to OpenSSL-based encryption (and this remains the default on Windows 7, which does not include the winSSL).

Click through for more information.

Comments closed

Scraping SQL Saturday Statistics

Published 2017-11-14 by Kevin Feasel

Tomaz Kastrun shows how to use rvest to read the SQL Saturday website and parse schedule details:

I wanted to check a simple query: How many times has a particular topic been presented and from how many different presenters.

Sounds interesting, tackling the problem should not be a problem, just that the end numbers may vary, since there will be some text analysis included.

Read on for the code and some analysis.

Comments closed

Offline Installation Of SQL Server 2017 ML Services

Published 2017-11-13 by Kevin Feasel

Jan Mulkens shows how to install SQL Server 2017 Machine Learning Services when your the server hosting SQL Server doesn’t have outbound internet access:

That’s when you remember it… Your server isn’t connected to the internet!
Pretty normal, but in your enthusiasm you completely forgot that SQL Server needs to download some binaries for the R and Python components you so desperately want on your precious machine!

Luckily, the installer comes to your rescue and shows you where to download those binaries it needs.
Turns out however… This link only is for one R component and the installer won’t let you pass to the next screen!

Read on for the answer.

Comments closed

Rendering Ten Million Points With ggplot2

Published 2017-11-08 by Kevin Feasel

Antonio Sánchez Chinchón shows how to draw Clifford attractors in R:

From a technical point of view, the challenge is creating a data frame with all locations, since it must have 10 milion rows and must be populated sequentially. A very fast way to do it is using Rcpp package. To render the plot I use ggplot, which works quite well. Here you have the code to play with Clifford Attractors if you want:

Click through for the code, as well as sample output images.

Comments closed

Multiple Result Sets With ML Services

Published 2017-11-07 by Kevin Feasel

Dave Mason figures out how to create multiple result sets with SQL Server ML Services:

Of course for this strategy to work, I’d have to know ahead of time how many data frames/HTML tables there are. Hmmm. Can dynamic T-SQL help me here? If I could find out at run time how many data frames there are, and which ones I may or may not want, then why not? Here’s some R code that reads HTML tables into a variable as a list of data frames(line 8), iterates through the list (starting at line 18), decides if the HTML table has any data in it (lines 21, 24), and adds the HTML table number (the element number in the list) to a different data frame (line 27). The output shows us we would want HTML tables 1, 2, and 4. (Yeah, I really didn’t want #4. But that can be fixed by enhancing the R code to be more selective. Let’s just go with it for now.)

The method is a bit disappointing (and it’s arguably worse for inputs); I do hope the ML Services team can improve upon this experience.

Comments closed