Leonardo Berbare de Araujo

Grad student of Computer Engineering at the University of São Paulo. Spends too much time with a bunch of TV, book and game series, but programming is always on his plans.

Read all articles published by Leonardo Berbare de Araujo.

Unsupervised learning shows which senators vote alike
Gravatar published this on

Data Science Programming

On this article, I get into details of how to write a program capable of grouping Brazilian senators according to their votes. All the way from getting the necessary info, going through some code and math, and finally getting to the Senate's inner circles.


Brazilian's legislative scenery is complex and varies now and then. Basic comprehension of how our representatives behave is usually challenging even to the most well-informed of us. To the ordinary citizen, this information may not arrive, and if it does, it's usually uninterpretable or too prepared.

Luckily, law enforces our government's transparency and accountability; that valuable information can be found on online portals like the one about the Senate. Don't get too excited, though, as information access is another right we've been struggling to wholly get in Brazil.

I decided to start this activity during internship at Infosimples in order to extract impartial yet meaningful conclusions from politicians' votes. It was worth the effort, since it put under test some concepts seen at Udacity's Intro to Artificial Intelligence (well, it also brought up curious readers like you! Leave some feedback later).


Read more about: