A portfolio analytics platform can implement a crawler and natural language processor to identify relevant articles. The natural language processor can integrate a text analysis tool, domain specific latent Dirichlet allocation tool, and theme measurement tool for identifying themes relevant to a particular domain. For example, the domain specific latent Dirichlet allocation tool identifies domain specific themes and uses an iterative process for eliminating articles or themes that are not specific to the domain. The theme measurement tool uses term frequency-inverse document frequency for naming and identifying the most relevant themes. The platform can generate interactive visual elements for an interface application.