<p>Thanks to the digitization of numerous texts, topic modeling methods are progressively used in various domains, like Digital Humanities. We propose to introduce formal concept analysis (FCA) as another exact method within the different families of topic modeling methods. FCA is a well-known paradigm based on the construction of a lattice that can be analyzed to find hidden relations or even calculate specific measures. In our case, we use mutual impact to visualize the main terms and the relevance of documents, and conceptual similarity to extract topics. To assess FCA as a topic modeling method, we present CREA, a text processing pipeline, and demonstrate its capabilities through a use case in which documents are analyzed to extract topics. Specifically, we reused multiple existing course materials to build a new course while identifying and isolating irrelevant documents within the corpus.</p>