Abstract: We introduce a scalable statistical platform to study non-causal or causal relationships of discrete attributes under network interference. The statistical platform is scalable in the sense that (a) it accommodates small and large populations with observed and unobserved heterogeneity and (b) it is amenable to scalable statistical computing. Scalable statistical computing relies on convex optimization based on pseudo-likelihoods, implemented by minorization-maximization algorithms. We establish theoretical guarantees based on a single observation of discrete and dependent attributes and connections and present simulation results along with an application to political discourse among U.S.\ state legislators on the social media platform X (Twitter).