- The
*R is my friend*blog publishes a series of four articles on neural networks. This is probably one of the most comprehensive introductions to neural networks in R. If you are in love with neural nets and want to learn even more, here is another tutorial by Saptarsi Goswami. - State-by-state media preferences as revealed by bit.ly.
- Andrew Gelman, Professor of Statistics and Political Sciences at Columbia University, discusses why Bing is preferred to Google by people who aren’t like him.
- Have you heard of Simpson’s Paradox? Here is an interactive visual (using the 1973 Berkeley sex discrimination lawsuit as an example) that explains the paradox in 60 seconds.
- Dan Delany does a visual breakdown of furloughed employees due to the U.S. government shutdown. The main view shows furloughed proportions by department, and there are real time tickers for duration, estimated unpaid salary, and estimated food vouchers unpaid.
- If there is an 82% chance an an event will occur within your life time (and assuming that you live for 70 years), what is the probability that this event will occur on any given day?
- Tableau, the popular interactive data visualization tool, is coming out with a new 8.1 update, and it will include integration with the R language. Learn how to integrate the two in just 30 seconds.
- A short (but not trivial) lesson on data smoothing using R.

## stats

14

Oct 13

## The week in stats (Oct. 14th edition)

7

Oct 13

## The week in stats (Oct. 7th edition)

- The picture above is a very well-known mathematical construction called the fractal cat. Brian Lee Yung Rowe shows how to construct fractal artworks using R.
- Arthur Charpentier of Freakonometrics explains how to construct ROC (
~~rate of change~~Receiver Operating Characteristic) curves in R, as well as how to interpret and plot them. This is a useful for those in fields that frequently encounter longitudinal data, such as finance, engineering or biostatistics. - There are many kinds of intervals in statistics. To name a few of the common ones: confidence intervals, prediction intervals, credible intervals, and tolerance intervals. Each are useful and serve their own purpose. You should not only know their names, but also when to use them and why.
- A map of the most visited website for every country in the world (source: Alexa.com), as well as the internet population of each country.
- Suppose that you drop 5 blue marbles and 5 red marbles randomly (and uniformly) on the interval [0,1]. What is the probability that the marbles will interleave each other?

30

Sep 13

## The week in stats (Sept. 30th edition)

- Given P(X = E(X)) = 1, does that mean Var(X) = 0?
- An interesting analysis of US high school graduation rates, conducted using R and googleVis.
- Do you have a unisex name? The following series of visuals tells us the most common unisex names in US history, and how the ratio of boys to girls changes over time.
- Most of us know what instrumental variables are (if not, here’s the Wikipedia page), but do you know what weak instruments are? The diffuseprior blog has a tutorial and tells you how to find them using R.

16

Sep 13

## The week in stats (Sept. 16th edition)

- This week, we found a number of useful webinars and presentations for statisticians and data scientists on R. Feel free to check out the following opportunities: Online course on forecasting using R by Prof. Hyndman of Monash University, Coursera’s free R courses, Why use R for Data Analysis by Vivek H. Patil of Gonzaga University, and two workshops on R by Bob Muenchen.
- If I roll five dice, what’s the chance that exactly two of the die show the same number?
- Did you know that even famous mathematicians like Paul Erdős had a hard time believing the result of the Monty Hall Problem? It was a computer simulation that eventually convinced him. Here’s a simulation of the Monty Hall Problem, and my own take on the how the problem is often poorly presented.
- During the 2013 JSM (Joint Statistics Meetings) Conference in Montreal, Revolution Analytics conducted a survey of attendees from August 5 to August 8. The 865 respondents gave their opinions on the privacy and ethics related to data collection, and on their familiarity with statistical software used for the analysis of such data. Out of the 865 statisticians surveyed…

9

Sep 13

## The week in stats (Sept. 9th edition)

- Larry Wasserman, Professor at Carnegie Mellon University, is a graduate of University of Toronto, a COPSS Award winner, and a leading statistician in Bayesian analysis and inference. In this post, he discusses his views on the question
*Is Bayesian Inference a Religion?* - Two people will each spend 15 consecutive minutes in a bar between 12:00pm – 1:00pm. Assuming uniform and independent arrival times, what is the probability that they will have a chance to clink glasses?
- Have you ever wondered which statistical package gives the fastest computational speeds? This quick comparison of Julia, Python, R and pqR provides some guidence.
- An interesting analysis of the most popular porn searches in the US.
- A quiz for everyone in the data visualization industry: Identify at least three problems with this chart and explain what you can do to make it better.
- R user groups continue to thrive worldwide. Joseph Rickert from Revolution Analytics prepares the following compilation of the locations of 127 R user groups around the world.