Profile cover photo
Profile photo
Mattias Villani
Natural Born Bayesian
Natural Born Bayesian

Mattias's posts

Post has shared content
Statistics in Google Sheets. Why? Because you can.
I'm happy to announce a new "Statistics" add-on for Google Sheets (the spreadsheet component of Google docs).

The add-on provides statistics and data analysis functionality right in Google Sheets, so you don't need to download your data to a separate customized statistics application. It is designed to get you a full statistical analysis of your data with very few clicks. You don't make a histogram, then make a boxplot, then compute the mean and standard deviation. Instead, you select the variables you want to analyze, and do the analysis all at once.

The app focuses on descriptive statistics and regression modeling. When paired with native spreadsheet support for hypothesis testing, this should be enough to cover the material taught in a 1-2 semester introductory statistics course. We plan to expand the modeling capabilities over time. Time series and generalized linear models are both obvious next steps.

We hope the add-on proves useful to both the education and enterprise domains. The aim is to allow analysts to actually use at work what they were taught at university. We have paid special attention to privacy and security. The back end computations are performed on secure Google servers, so your data never leaves the Google ecosystem. You can rely on the Google Sheets permission system to decide who can see your data and your analysis.

You can install the app directly from the Google Sheets menu.

Add-Ons ->
Get Add-Ons ->
Search for "Statistics"
We're the one from "Google Statisticians and Engineers"

Or you can follow the installation instructions on page linked to by this post.

Finally, everyone involved in creating this app is a Googler, but it is not an official Google product, so we don't have the Google marketing machine at our disposal. We need some help getting the word out. If you know someone who might be interested in this, please tweet, blog, or post about us to your favorite social network. Thanks!

Post has attachment
Another one bites the dust.

Post has attachment
From the classic paper "In Defense of Probability" by Peter Cheeseman:

"In this paper, it is argued that probability theory, when
used correctly, is sufficient for the task of reasoning under
uncertainty. Since numerous authors have rejected probability
as inadequate for various reasons, the bulk of the
paper is aimed at refuting these claims and indicating the
sources of error. In particular, the definition of probability
as a measure of belief rather than a frequency ratio is advocated,
since a frequency interpretation of probability drastically
restricts the domain of applicability."

Post has attachment
The Division of Statistics and Machine Learning at Linköping University in Sweden are looking for senior lecturers in Statistics with 50% research time for at least 5 years.

Our division is quite unique in Sweden in that we belong to Computer Science department and we take the interaction between Statistics and Computer Science quite seriously. For example, we lead the Machine Learning Research Group at Linköping University (

We host an international master's programme 'Statistics and Data Mining' which is taught completely in English, and the largest Swedish undergraduate programme in Statistics, 'Statistisk och Datanalys'.

More details about the positions can be found here:

Deadline for applications: February, 23.

Post has shared content
Judging from Google+ posts, this has to be the most dynamic and fun group in the UK. 
Delighted to be advertising for a Chair in Operational Research to join our thriving and growing group!  

Post has attachment
Just got myself a Moto 360. A nice creative feature is that one can design watchfaces. I couldn't resist. 

Post has attachment
Christmas gift from nvidia: 13000 USD worth of graphics cards for high performance GPU computing. I must have been good this year. 

Post has attachment
A word from our sponsors ...

Post has attachment
Very clear exposition of Bayesian model comparison (hypothesis testing). Clearly explains why the prior is so important for this type of analysis. Such a simple message, but very often lost on practitioners of Bayesian model comparison. 
Wait while more posts are being loaded