politics · internet

## Hackers, datos y civismo

En las últimas semanas he vuelto, de una u otra forma, a pensar sobre los datathons (hackathons especializados en la explotación de datos) como laboratorios de ideas y en su utilidad para mejorar las políticas públicas urbanas y la interacción entre ciudadanía y policy makers. Las iniciativas que he visto recientemente tienen una vocación de permanencia diferente a la de los eventos puntuales, como el muy celebrado Apps for Democracy, pero están imbuidos, si es que no acentúan, el mismo optimismo que ha venido caracterizando a toda la revolución de los datos. Por fortuna, estas propuestas de más larga duración... Read more ›

music · random

## Puccini

It seems that the landmark that signals that one has abandoned the neophite stage in opera happens when one starts to feel contempt for Puccini. There is no shortage of irony in the fact that most of us started listening to opera precisely because of him, but it is true that his operas soon feel very unsophisticated, too saccharine. And yet, I felt defensive when Anthony Arblaster goes out of his way to single out Puccini and Richard Wagner in Viva La Libertà!: Politics in Opera for not writing about politics. Or even worse, because politics happens in the background,... Read more ›

julia · stats

## First steps with Julia

I went back to Julia this week. I took a very cursory look at the documentation when I was entering the final leg of my dissertation a couple of years ago but I did not pay much attention to the project –learning another language was the last thing on my mind back then. However, this week I decided that it was about time to give it a proper try: I have been wanting to expand my toolset and Julia seems as good a candidate as any other, specially given its growing popularity and its impecable credentials. Because the language is... Read more ›

stats

## Optimal sample allocation in experiments

$$\newcommand{\trace}{\mathop{\rm trace}\nolimits}$$ One of the recurrent problems I face at work is about sample allocation. The situation is as follows. Given a fixed sample size that is determined by an external “budget,” I have to perform an experiment to test a standard treatment against two alternatives to check which one is superior. For instance, we have a feature in place that I need to test against two possible replacements to know which one we should use moving forward. Because of that, I am not interested in the value of each treatment, as much as on the difference between... Read more ›

rstats · stats

## Likelihood and entropy. An example

As an example of the ideas in the previous post, I am running here an example of an empirical likelihood estimation in the simplest possible setting following this example. Asumme n=50 observations is extracted from $$x_i \sim N(4, 1), i=1,…,n: N &lt;- 50 x &lt;- rnorm(N, 4, 1) The maximum likelihood estimate of the mean is simply print(mean(x)) ## [1] 4.086 Maximizing the program in the post with respect to$$\lambda_2\$1, resuts in the following condition and we can find the zeroes of the previous implicit function by minimizing: lambda &lt;- function(ll, x, mu) { n &lt;- length(x) num &lt;-... Read more ›