The effects of the 2007 financial crisis in the scope of housing access and loss in Spain have been devastating for a great part of the population with consequences still visible nowadays. Data available unveils the ineffectiveness of the approach undertaken by legislators and the judiciary in order to confront the complex situation. We present a methodology that aims at studying the jurisprudence by scrutinizing data from judicial decisions so to reveal systemic patterns, pillar decisions and the consequences of important legislative changes, and thus be able to evaluate those mechanisms that fail along the judiciary process.
Using a corpus of housing-related decisions that spans over the last 20 years, we focus on their textual content. We characterize each text as a distribution over topics. To do so, we make use of a network approach [1] that infers a stochastic block model [2] from the bipartite network formed by the corpus of documents and their words. Then, the blocks or communities of words detected by this procedure serve as topics for each document.
Then, once each text is expressed a weighted collection of topics, we aggregate the results over the entire corpus so to analyze the historical discursive evolution, thus finding trends and important changes. Specifically, by means of information-theory based measures of divergence [3], we are able to quantify not only the change but also the propagation of textual content in time.
Preliminary results show a major change in the textual content of decisions some time after the 2007 financial crisis, showing the extent to which the judiciary responded to the critical situation. This methodology shows that we are able to detect important social changes through historical discursive changes in the jurisprudence.