The R Backpages

by Joseph Rickert

As an avid newspaper reader (I still get the print edition of the New York Times delivered every Sunday morning) I have always thought that some of the most interesting news is to be found in the back pages. So, in that spirit here are some things that I thought might be fit to print.

Plotly has new R interface in Beta Test

Plotly, a startup building a data analysis and visualation collaboration site, is beta testing their new R library and API and the team there is looking for feedback. Writing to Revolution Analytics' community site, Matt Sundquist, one of Plotly's cofounders, is asking for feedback, suggestions for the library, and tweets (@plotlygraphs) to help get the word out.

To get started with plot.ly, first sign up for an account and an autentication key. Install the the plotly library from GitHub (The site provides clear instructions and several examples). Then write some code in your local R environment and let the plotly API takes care of interacting with the website. The following code produces the boxplots below which can be viewed by going to the web page returned in the variable url.

#Box plot Example
library(plotly)
username="RevoJoe"
api_key = "RevoJoe's key"
p <- plotly(username=username, key=api_key)
box1 = list(y=c(0,1,2,4),type='box')
box2 = list(y=c(1,2,4,8),type='box') 
response = p$plotly(list(box1,box2))
url = response$url
filename = response$filename

The web page is interactive enabling users to get information about the data by mousing over portions of portions of the plot, to add titles and further refine the plots, and then share them via browser interface. This graph shows government shutdown percentages, by agency. Hardest hit were: NASA, the Department of Housing and Urban Development, the DOE, and the EPA.

New Book on R Visualizations

In their upcoming meeting the Köln R User Group will be looking into the new book "Datendesign mit R" by Thomas Rahlm. The website for the book shows some very impressive examples of visualizing data with R which are backed by very readable code.

Live Meetup Streaming from Munich

The Munich useR Group has begun streaming their meetings live. Markus Schmidberger reports that "the live stream has a delay for about 3-5 seconds and the quality is perfect". Because the Munich group regularly hosts high quality talks this innovaion should be of interest to the worldwide R community. You can view past talks on the group's YouTube channel.

Quandl Adds More Data

Quandl, the R Friendly source for data, continues to add to trove of over 7,000,000 data sets. Recent arrivals include:

• Updated Commodities Page with spot and/or futures prices for 86 physical commodities and 12 commodity indexes from around the world
• Detailed commodity data pages for: Metals, Natural Gas, Grains and Cereals, & Farms and Fishery
• New country topic pages including economic, demographic, society data and more for:India, Russia, and Indonesia
• National Institute of Statistics France: 110,000 datasets on the French economy including national accounts, price indices, employment, and more
• India Ministry of Statistics: 3,000 datasets on all aspects of Indian society, such as economic indicators, finance, transportation, agriculture, labor, and more
• India Department of Commerce: Monthly import and export data from India going back 20 years
• Freddie Mac: 400 housing price indices from different sized geographical areas all over the United States
• Policy Uncertainty Project: Economic Policy Uncertainty Indices for the U.S., Canada, India, China and the EU.
• European Central Bank: 17,000 datasets on various economic indicators primarily related to countries within the European Union.

Webinars

• The Toronto R Matlab users does webinars in addition to their local meetups. The next one on "Live Trading Strategy with Execution via Interactive Brokers" will be held on November 19th. 
• Lion Systems is offering a free webinar on "R for SAS and SPSS" users today (10AM PST).
• Revolution Analytics has a whole series of webinars that may be viewed online.

Scheduled Talk

I am looking forward to giving a Tech Talk on doing statistics with large data sets at the Predictive Analytics Center of Excellence, PACE, on November 13th in San Diego.

If you have some news that you think should be in a future edition of the R Backpages please let me know.