We use natural language processing techniques to explore issue attention in the decisions of the European Court of Justice. Analysing the entire universe of ECJ decisions up to 2015, we map issue prevalence across time, procedures and litigant type. We find evidence that the more inclusive annulment and referral procedures are associated with greater issue heterogeneity whereas less inclusive infringement procedure displays greater issue cohesiveness as well as greater issue stability over time. The saliency of internal market and environmental issues in infringement proceedings is, we argue, a direct consequence of the European Commission's tight grip over the Court's agenda. Interestingly, while preliminary rulings also emphasise internal market themes, they devote far more attention to employment, social and immigration questions. We use a citation network constructed from a qualitative study of family reunion cases to demonstrate the validity of our computerised case classification. We conclude that text-mining methods represent a powerful and reliable tool to investigate judicial agenda formation on international courts.