Ministry of Social Development, Mexico
Complex Network Analysis of Political Data The Mexican Senate was recently awarded for fulfilling its transparency obligations, meaning that a vast amount of information is on their website. The problem is that the way in which they present it is largely uninformational so a complex network approach is proposed to study the dynamics of the given system. We use Cascadia package and Julia’s parallelization capabilities in order to scrap the data that we consider useful for our research and export it to the graph oriented database Neo4j and make an exploratory data analysis for attendancy and voting patterns. Using LightGraphs package we build a temporal weighted network and apply a self developed algorithm for community detection in order to study coalition dynamics. Finally we make use of RCall to build a visualization that represents such dynamics through an alluvial diagram getting interest insight such as party switching and opposite parties coalitions. As we want to make this as automatic and portable as possible, everything runs in Docker containers and is orchestrated via Docker Compose and a Luigi Pipeline. That gives us the opportunity to enlarge and analyze new information automatically once a week.
Musician who studied Physics and Data Science. Currently working at the Ministry of Social Development in Mexico using data science to focalize aid to those with greater needs.
