Revolutions

A Certification for R Package Quality

There are more than 12,000 packages for R available on CRAN, and many others available on Github and elsewhere. But how can you be sure that a given R package follows best development practices for high-quality, secure software?

Based on a recent survey of R users related to challenges in selecting R packages, the R Consortium now recommends a way for package authors to self-validate that their package follows best practices for development. The CII Best Practices Badge Program, developed by the Linux Foundation's Core Infrastructure Initiative, defines a set of criteria that open-source software projects should follow for quality and security. The criteria relate to open-source license standards, documentation, secure development and delivery practices, version control and bug reporting practices, build and test processes, and much more. R packages that meet these standards are a signal to R users that can be relied upon, and as the standards document notes:

There is no set of practices that can guarantee that software will never have defects or vulnerabilities; even formal methods can fail if the specifications or assumptions are wrong. Nor is there any set of practices that can guarantee that a project will sustain a healthy and well-functioning development community. However, following best practices can help improve the results of projects. For example, some practices enable multi-person review before release, which can both help find otherwise hard-to-find technical vulnerabilities and help build trust and a desire for repeated interaction among developers from different organizations.

Developers can self-validate their packages (free of charge) using an online tool to check their adherence to the standards and get a "passing", "silver" or "gold" rating. Passing projects are eligible to display the badge (shown above) as a sign of quality and security, and almost 200 projects qualify as of this writing. A number of R packages have already gone through or started the certification process, including ggplot2, R Consortium projects covr and DBI, and the built-in R package Matrix Matrix. For more details on how R packages can get to a passing certification, and the R Consortium survey they led to the recommendation, see the R Consortium blog post at the link below.

R Consortium: Should R Consortium Recommend CII Best Practices Badge for R Packages: Latest Survey Results

 

Because it's Friday: Street Orientation

Most cities in the US have a grid-based street structure. But it's rarely a perfect grid: sometimes the vagaries of history, geography, or convenience lead to deviations from right angles. And sometimes, rival urban planners simply disagree:

When two city planners hate each other: pic.twitter.com/LSb7k8KoaW

— Wilson 🏳️‍🌈🌹 (@the_sidecarist) March 8, 2018

Geoff Boeing, an urban planning postdoc at UC Berkeley, developed an interesting way to visualize the "griddedness" of cities, by summarizing at the distribution of street segments in a polar plot (you can see the Python code behind the charts here):

I was surprised to see Seattle with such a regular north-south/east-west distribution, whereas in fact each Seattle neighborhood has its one grid aligned to its own orientation (often dictated by the shape of the local coastline). But of course the polar chart depends on how much of the city you include, and I suppose the one above includes just the downtown streets. But you can create your own charts using any zone you like, using this online tool by Randy Olson:

That's all for the blog for this week. Have a great weekend, and we'll be back on Monday. See you then!

 

aRt with code

Looking for something original to decorate your wall? Art With Code, created by Harvard University bioinformatician Jean Fan, provides a collection of R scripts to generate artistic images in the style of famous artworks, for example this randomly-generated piece in the style of Mondrian:

Other art generators include "Tunnel" (rotated and scaled designs in the style of Päivi Julin), Pointillism (sylizes photographs as a dots), and Pixel (an interesting multi-scale filter for photos). The transformations using only the built-in R grid and lattice packages, which made it easy for me to try Pixel on a photo of my own dog, Max:

You can find examples of all of these styles and the R scripts used to generate them at the link below.

Jean Fan: Art With Code  (via the author)

A quick tour of AI services in Azure

If you're after a quick overview of some of the services available in Azure to build AI-enabled applications, you might want to check out the 6-minute video below. It provides a tour of three services:

• Azure Cognitive Services, pre-trained machine learning models to detect and characterize faces, text sentiment, and more 
• Azure Machine Learning, a development, deployment and management framework for machine learning models you train yourself, and
• Azure Bot Service, a framework for creating intelligent chat-based interfaces.

Want to learn more? You can find more videos and resources at the Azure Essentials page.

AI, Machine Learning and Data Science Roundup: July 2018

A monthly roundup of news about Artificial Intelligence, Machine Learning and Data Science. This is an eclectic collection of interesting blog posts, software announcements and data applications I've noted over the past month or so.

Open Source AI, ML & Data Science News

A review of the current state of the Julia project, including performance comparisons with Go, Python and R. 

Python 3.7.0 has been released. Also, Guido van Rossum steps down as the project's "Benevolent Dictator for Life".

Tensorflow 1.9 released, featuring a new getting started guide for Keras.

Apache OpenNLP 1.9.0 released, adding several improvements to the natural language processing library.

The PYPL Popularity of Languages Index ranks Python as the 1st and R as the 7th most popular programming languages.

R 3.5.1 has been released.

Industry News

Google introduces Cirq, an open-source framework to simulate noisy intermediate-scale quantum computers.

Hortonworks announces new partnerships with Google, Microsoft and IBM to bring its big data platform to the cloud.

IBM plans to release an annotated dataset of 1 million images, with the goal of reducing bias in facial recognition systems.

Amazon Sagemaker enhances the DeepAR, BlazingText, and Linear Learner algorithms built into the machine learning framework.

A series on Data Ethics by Mike Loukides, Hilary Mason and DJ Patil. The most recent post proposes a checklist for ethical data science practices.

Apple has combined its various artificial intelligence divisions into a new structure led by recent hire John Giannandrea.

Microsoft News

Open data sets curated by Microsoft Research for computer vision, natural language processing and scientific analysis. 

Microsoft acquires Bonsai, a startup focused on using ML to train autonomous machines using simulated environments.

Microsoft President Brad Smith calls for public regulation and corporate responsibility in the use of facial recognition technology.

ML.NET 0.3 released adding ONNX support, LightGBM learners, and more to the .NET machine learning library.

Microsoft Speech Service SDK version 0.5.0 has been released, with support for long-running audio and automatic reconnections.

Azure HDInsight now supports Apache Spark 2.3, improving performance for Python-based interactions.

Azure Databricks now includes RStudio integration for R developers. The integration is described in detail at the Databricks blog.

The Microsoft Face API has significantly improved its ability to recognize gender across skin tones.

Learning resources

Andrey Kurenkov's essay on the limitations of reinforcement learning. A February 2018 blog post by Alex Irpan also provides some cautionary examples.

A Developer's Guide to Building AI Applications, a free e-book that provides a gentle introduction to the Microsoft AI Platform.

A tutorial on adding machine learning to mixed reality applications, using Azure ML Studio.

How to read and write Microsoft SQL Server data in Python using Pandas dataframes.

Object dectection via distributed deep learning on multiple GPUs, using Horovid on Azure Batch AI.

How Artificial Intelligence Could Prevent Natural Disasters, from Wired.

A neural network for colorizing video also finds application in object tracking.

Computer vision with an embedded learning library from Microsoft Research used to monitor endangered Purple Martin nests at Disney Animal World.

A diabetic retinopathy prediction application, based on image analysis with convolutional neural networks and implemented with Azure Machine Learning.

Fashion similarity matching with Siamese Networks.

Find previous editions of the monthly AI roundup here

A hex sticker wall, created with R

Bill Venables, member of the R Foundation, co-author of the Introduction to R manual, R package developer, and one of the smartest and nicest (a rare combo!) people you will ever meet, received some gifts from the user!2018 conference committee in recognition of his longtime service to the R community. He received a book of reminiscences from those whose lives he has touched, and this blanket:

This morning we took care of some unfinished #user2018 business. 😊 @visnut @tslumley @_jessie_roberts pic.twitter.com/gW0wX2PLF2

— Miles McBain (@MilesMcBain) July 19, 2018

The design on the blanket is the "Hex Sticker Wall" that was on design during the conference, created by Mitchell O'Hara-Wild:

The #useR2018 #hexwall has been revealed! Read about how it was created in #rstats on this blog post: https://t.co/krYYOQ3N84

A huge thanks to everyone who has submitted stickers and provided feedback. I hope you enjoy the end result as much as I have had creating it! 🎉 pic.twitter.com/GnG9m2cZme

— Mitchell O'Hara-Wild (@mitchoharawild) July 11, 2018

The design for the wall was created — naturally — using R. Mitchell created an R script that will arrange a folder of hexagon images according to a specified layout. He then used a map of Australia to lay out the hex sticker coordinates within the map boundaries and plotted that to create the Hexwall. If you have a similar collection of hex images you could use the same technique to arrange them into any shape you like. The details are provided in the blog post linked below.

Mitchell O'Hara-Wild: UseR! 2018 Feature Wall

Highlights from the useR! 2018 conference in Brisbane

The fourteenth annual worldwide R user conference, useR!2018, was held last week in Brisbane, Australia and it was an outstanding success. The conference attracted around 600 users from around the world and — as the first held in the Southern hemisphere — brought many first-time conference-goers to useR!. (There were also a number of beginning R users as well, judging from the attendance at the beginner's tutorial hosted by R-Ladies.) The program included 19 3-hour workshops, 6 keynote presentations, and more than 200 contributed talks, lightning talks and posters on using, extending, and deploying R.

If you weren't able to make it to Brisbane, you can nonetheless relive the experience thanks to the recorded videos. Almost all of the tutorials, keynotes and talks are available to view for free, courtesy of the R Consortium. (A few remain to be posted, so keep an eye on the channel.) Here are a few of my personal highlights, based on talks I saw in Brisbane or have managed to catch online since then.

Keynote talks

Steph de Silva, Beyond Syntax: on the power and potentiality of deep open source communities. A moving look at how open source communities, and especially R, grow and evolve.

Bill Venables, Adventures with R. It was wonderful to see the story and details behind an elegantly designed experiment investigating spoken language, and this example was used to great effect to contrast the definitions of "Statistics" and "Data Science". Bill also includes the best piece advice to give anyone joining a specialized group: "Everyone here is smart; distinguish yourself by being kind".

Kelly O'Brian's short history of RStudio was an interesting look at the impact of RStudio (the IDE and the company) on the R ecosystem.   

Thomas Lin Pedersen, The Grammar of Graphics. A really thought-provoking talk about the place of animations in the sphere of data visualization, and an introduction to the gganimate package which extends ggplot2 in a really elegant and powerful way.

Danielle Navarro, R for Pysychological Science. A great case study in introducing statistical programming to social scientists.

Roger Peng, Teaching R to New Users. A fascinating history of the R project, and how changes in the user community have been reflected in changes in programming frameworks. The companion essay summarizes the talk clearly and concisely.

Jenny Bryan, Code Smells and Feels. This was an amazing talk with practical recommendations for better R coding practices. The slides are available to view online.

Contributed talks

Bryan Galvin, Moving from Prototype to Production in R, a look inside the machine learning infrastructure at Netflix. Who says R doesn't scale?

Peter Dalgaard, What's in a Name? The secrets of the R build and release process, and the story behind their codenames.

Martin Maechler, Helping R to be (even more) Accurate. On R's near-obsessive attention to the details of computational accuracy.

Rob Hyndman, Tidy Forecasting in R. The next generation of time series forecasting methods in R. 

Nicholas Tierney, Maxcovr: Find the best locations for facilities using the maximal covering location problem. Giftastic!

David Smith Speeding up computations in R with parallel programming in the cloud. My talk on the doAzureParallel package.

David Smith, The Voice of the R Community. My talk for the R Consortium with the results of their community survey.

In addition, several of my colleagues from Microsoft were in attendance (Microsoft was a proud Platinum sponsor of useR!2018) and delivered talks of their own:

Hong Ooi, SAR: a practical, rating-free hybrid recommender for large data

Angus Taylor, Deep Learning at Scale with Azure Batch AI

Miguel Fierro, Spark on Demand with AZTK

Fang Zhou, Jumpstart Machine Learning with Pre-Trained Models

Le Zhang, Build scalable Shiny applications for employee attrition prediction on Azure cloud

Looking back, looking ahead

Overall, I thought useR!2018 was a wonderful conference. Great talks, friendly people, and impeccably organized. Kudos to all of the organizing committee, and particularly Di Cook, for putting together such a fantastic event. Next year's conference will be held in Toulouse, France and already has a great set of keynote speakers announced. But in the meantime, you can catch up on the talks from useR!2018 at the R Consortium YouTube channel linked below.

YouTube: R Consortium

Video: R for AI, and the Not Hotdog workshop

Earlier this year at the QCon.ai conference, I gave a short presentation, "The Case for R, for AI developers". I also presented an interactive workshop, using R and the Microsoft Cognitive Services APIs, to automatically generate captions for images, and to create a tool to recognize images of hotdogs. Video from both the presentation, and the workshop (which starts at the 10:00 mark), is now available to view the QCon.ai website.

You can find the slides for the presentation here. The R code for the "Not Hotdog" workshop is available as an Azure Notebook which you can clone and use to follow along with the workshop.

InfoQ: R for AI developers

Because it's Friday: Language and Thought

Does the language we speak change the way we think? This TED talk by Lera Boroditsky looks at how language structures like gendered nouns, or the way directions are described, might shape they way speakers of those languages think about things:

This talk was cited by Bill Venables in his excellent keynote talk at the useR!2018 conference, where he also presented the results of an elegant designed experiment that looked at the influence of gendered nouns by comparing languages without gendered nouns (like Hungarian) with those that have multiple genders for nouns like Spanish (which has feminine and masculine nouns) and German (which adds neuter nouns, for three genders total).

That's all from us for this week. Have a great weekend, and we'll be back next week. Enjoy!

New open data sets from Microsoft Research

Microsoft has released a number of data sets produced by Microsoft Research and made them available for download at Microsoft Research Open Data.  

The Datasets in Microsoft Research Open Data are categorized by their primary research area, such as Physics, Social Science, Environmental Science, and Information Science. Many of the data sets have not been previously available to the public, and many are large and useful for research in AI and Machine Learning techniques. Many of the datasets also include links to associated papers from Microsoft Research. For example, the 10Gb DESM Word Embeddings dataset provides the IN and the OUT word2vec embeddings for 2.7M words trained on a Bing query corpus of 600M+ queries.

Other data sets of note include:

• A collection of 38M tweets related to the 2012 US election
• 3-D capture data from individuals performing a variety of hand gestures
• Infer.NET, a framework for running Bayesian inference in graphical models
• Images for 1 million celebrities, and associated tags
• MS MARCO, is a new large-scale dataset for reading comprehension and question answering

Most data sets are provided as plain text files, suitable for importing into Python, R, or other analysis tools. In addition to downloading the data, you can also deploy the datasets for analysis in to Microsoft Azure: see the FAQ for details on this and other ways the datasets can be used.

As of this writing, the archive includes 51 datasets. You can explore and download the Microsoft Research datasets at the link below.

Microsoft: Microsoft Research Open Data