This paper investigates how the topics of European Union (EU) legislative proposals are correlated with political contestation within the Council of the EU. Because official policy areas are often too broad to capture meaningful differences between proposals, we classify each legislative initiative using the EuroVoc thesaurus terms assigned to it, creating a detailed categorization that identifies both individual topics and combinations of topics. This approach reflects the reality that many proposals span multiple issue areas and allows us to examine how different thematic profiles correlate with disagreement among member states. By linking these text-based classifications to measures of opposition in Council negotiations, we identify which topics and which combinations of topics are correlated with higher levels of contestation. The results reveal nuanced patterns of politicization that standard classifications fail to capture. Methodologically, the study demonstrates the value of integrating automated text analysis with data on member state voting behaviour, while substantively it advances understanding of the dynamics of controversy in EU decision-making and provides a dataset that can support further research on legislative politics in the Council.