A vast amount of our cultural heritage is at risk of getting lost because it resides in old books that are difficult to access. It is therefore important to make this information available to human readers but also to machine analysis, so that new representations and insights based on this knowledge can be constructed. In our case study, we use a host of digital tools to extract and analyze a corpus of 19th century French texts about the practices of market gardening in Paris, and to apply a variety of possible visualizations in an integrated interface. Our work includes a Named Entity and Linking procedure for creating maps of the locations mentioned in these texts as well as the social networks of people cited in the books. We also consider how the analysis of verbs can approximate and represent the know-how of market gardening: we analyze the statistics of those verbs compared to their usage in a general corpus for French, and map the verbs using word embeddings. Finally, we also consider a semantic frame analysis to extract causal relations from texts to evaluate how well these relations support the biological knowledge embedded in those texts (such as how too much exposure to the sun may affect the quality of the garden’s produce). Altogether, we show how the visualizations based on Natural Language Processing and Textual Statistics could support a convivial navigation through the corpus.