For my PhD thesis titled The Emergence of Legislative Organization in the Colombian Congress (rice.edu) I explored the use of graph databases and network science to predict voting patterns in the Colombian Congress. The Colombian Legislative context is one where traditional center-left and center-right parties have splintered into smaller fractions, leading to less cohesive voting behavior and weaker party discipline. The Colombian electoral system creates incentives for legislators to focus on their personal reputations and not the party label, leading to the rise of independent candidates and weakening centralized party control over legislators. For this reason, traditional theories to predict voting behavior are not well-suited.

In my dissertation, I built on insights from complexity science to create a novel theory to explain the basis of voting coalitions. I built a graph database from multiple data sources and utilized network science to build models to uncover what accounts for voting behavior in this complex setting.