This project is meant to be used as a resource for studying language dynamics between different groups in society, focusing how they talk about immigration, ethnic minorities, and neighborhoods associated with high clustering of ethnic similar others, but also other things. How people talk about events, people and places are theorized to reflect their interpretations and beliefs about them as well. However, the empirical evidence of how people's perceptions and actions go together is still underdeveloped. In a series of paper, we will explore: (1) studying how language evolves over time when discussing different issues online, in academia, in newspapers, and in business. Answering the questions: what makes discussions evolve over time, and when do different groups in society converge/diverge in how they talk? What dynamics explains these patterns? (2) How well do measures of how people talk about places/groups correspond with the characteristics of these places/groups align? What can explain alignment, and does it matter for individual-level actions? We conduct (1) by using dynamic contextual word embeddings on different corpora and explore how different concepts are talk about in different contexts over time. We look at divergence and convergence by looking at how contextualized embeddings cluster in vector space over time, and how the clustering correspond to meaningful social categories. We explore the role of individual actors in producing the observed dynamics. We conduct (2) by comparing the context in which different groups of society discuss different groups and places and compare it to what we can learn about these groups and places from Swedish register data. Comparing how concepts relate in the language (using word embeddings) with how they relate in "reality" will help us answer long-standing questions about how people form their perceptions about the world. Understanding the diffusion of new language helps us understand the role of opinion homophily in language use and the politicizing power of languages.