distracting redistricting?

“We at FiveThirtyEight will be tracking the whole redistricting process, from proposed maps to final maps, so watch this space for updates!”

FiveThirtyEight is keeping a tracker on the “redistricting” of U.S. states, namely the decennial redrawing of electoral districts. This is still an early stage when no map has been validated by the state legislature and hence I cannot tell whether or not FiveThirtyEight will be analysing gerrymandering in a statistical manner, to figure out how extreme the map is within the collection of all electoral maps. The States being the States, the rules vary widely between them, from the legislators themselves setting the boundaries (while sometimes being very open on their intentions to favour their own side) to independent commissions being in charge. I did not spot any clear involvement of statisticians in the process.

“The application of differential privacy will bring significant harm to Alabama (…) The Census Bureau has not shown that other disclosure avoidance methods
would not satisfy the privacy requirements” Case No. 3:21-cv-00211

While looking at this highly informative webpage maintained by University of Colorado Law School Doug Spencer, I came across this federal court challenge by the State of Alabama again the Census Bureau for using differential privacy! A statistical version of “shoot the messenger”?! The legal argument of the State is “the Fifth Amendment, alleging that differential privacy is a violation of the one-person, one-vote principle and will result in the dilution of their votes.” I however wonder what is the genuine (political) reason for this challenge!

Adam Ondra completes a first 9c

In my office hangs this poster of Adam Ondra climbing at sunset an impressive overhang with a little Czech town at the foot of the cliff, already in the shade. Impressive because of the view and of the climb which at 7c is a whole grade (and then some) beyond my reach. But now Ondra has managed to climb the first 9c in the world, which is universes beyond the impressive and beyond the fathomable, with passages only manageable feet first. Which actually makes a lot of sense, the way he explains it. The route is currently called Project Hard and is located in Hanshelleren Cave, Flatanger, Norway.

Texan black swan

“Un événement improbable aux conséquences d’autant plus désastreuses que l’on ne s’y est pas préparé.”

This weekend, there was a short article in Le Monde about the Harvey storm as a Texan illustration of Taleb’s black swan. An analysis that would imply every extreme event like this “once-in-a-thousand year” event (?) can be called a black swan… “An improbable event with catastrophic consequences, the more because it had not been provisioned”, as the above quote translates. Ironically, there is another article in the same journal, about the catastrophe being “ordinary” and “not unexpected”! While such massive floods are indeed impacting a huge number of people and companies, because the storm happened to pour an unusual amount of rain right on top of Houston, they indeed remain within the predictable and not so improbable in terms of the amount of water deposited in the area and in terms of damages, given the amount and style of construction over flood plains. For instance, Houston is less than 50 feet above sea level, has fairly old drainage and pipe systems, and lacks a zoning code. With mostly one or two-story high buildings rather than higher rises. (Incidentally, I appreciated the juxtaposition of the article with the add for Le Monde des Religions and its picture of a devilesque black goat!)

 

Statistics month in Marseilles (CIRM)

Next February, the fabulous Centre International de Recherche en Mathématiques (CIRM) in Marseilles, France, will hold a Statistics month, with the following programme over five weeks

• 1 – 6 February, 2016  Learning
• 8 – 12 February, 2016 Mathématical statistics
• 15 – 19 February, 2016 Processes
• 22 – 26 February, 2016 Extremes, Copulas and Actuarial Science
• 29 February –  4 March, 2016 Bayesian statistics and algorithms 

Each week will see minicourses of a few hours (2-3) and advanced talks, leaving time for interactions and collaborations. (I will give one of those minicourses on Bayesian foundations.) The scientific organisers of the B’ week are Gilles Celeux and Nicolas Chopin.

The CIRM is a wonderful meeting place, in the mountains between Marseilles and Cassis, with many trails to walk and run, and hundreds of fantastic climbing routes in the Calanques at all levels. (In February, the sea is too cold to contemplate swimming. The good side is that it is not too warm to climb and the risk of bush fire is very low!) We stayed there with Jean-Michel Marin a few years ago when preparing Bayesian Essentials. The maths and stats library is well-provided, with permanent access for quiet working sessions. This is the French version of the equally fantastic German Mathematik Forschungsinstitut Oberwolfach. There will be financial support available from the supporting societies and research bodies, at least for young participants and the costs if any are low, for excellent food and excellent lodging. Definitely not a scam conference!

dynamic mixtures [at NBBC15]

A funny coincidence: as I was sitting next to Arnoldo Frigessi at the NBBC15 conference, I came upon a new question on Cross Validated about a dynamic mixture model he had developed in 2002 with Olga Haug and Håvård Rue [whom I also saw last week in Valencià]. The dynamic mixture model they proposed replaces the standard weights in the mixture with cumulative distribution functions, hence the term dynamic. Here is the version used in their paper (x>0)

where f is a Weibull density, g a generalised Pareto density, and w is the cdf of a Cauchy distribution [all distributions being endowed with standard parameters]. While the above object is not a mixture of a generalised Pareto and of a Weibull distributions (instead, it is a mixture of two non-standard distributions with unknown weights), it is close to the Weibull when x is near zero and ends up with the Pareto tail (when x is large). The question was about simulating from this distribution and, while an answer was in the paper, I replied on Cross Validated with an alternative accept-reject proposal and with a somewhat (if mildly) non-standard MCMC implementation enjoying a much higher acceptance rate and the same fit.