Size of circles

[subratac]

from the original paper of PyLDAviz (https://www.aclw'eb.org/anthology/W14-3110/), I understand the size (area) of the circle corresponds to the prevalence of the topic in the corpus which I am interpreting as the number of documents that are mapped to this topic. By mapping I mean by taking the document-topic matrix and assigning each document to the topic that has the maximum probability. Is my understanding correct?