Fig. 4

Analysis based on the semantic content of unstructured written reports. A. Networks constructed based on the semantic similarity of the reports associated with cultivars. The left panel is color-coded based on the results of modularity maximization using the Louvain algorithm, while the right panel is color-coded based on species tag. B. Word clouds representing the most frequent terms appearing in the reports of all cultivars combined (left), “sativas” (middle) and “indicas” (right)