A Dungeons and Dragons Data Set

This paper introduces the Forgotten Realms Wiki (FRW) data set and related analyses. Forgotten Realms is the de-facto default setting of the popular open-ended tabletop fantasy role-playing game, Dungeons & Dragons. The data set was extracted from the Forgotten Realms Fandom wiki consisting of more than 40,000 articles. The FRW data set is constituted of 11 sub-data sets in varying degrees of pre-processing: raw plain text (FRW-P), each article annotated by the original wiki article title (FRW-J), first paragraph plain text annotated by the original wiki article title (FRW-FJ), directional graphs with all links and first link (FRW-L and FRW-FL), directional graph with wiki category (FRW-CL), wiki info-boxes annotated by the original wiki article title (FRW-I), Poincar ´e embedding of FRW-FL (FRW-PE), Word2Vec models of FRW-P (FRW-W), Doc2Vec models of FRW-P and FRW-FJ (FRW-D and FRW-FD). This is the first data set of this size for the Dungeons & Dragons domain. We then present a pairwise similarity comparison benchmark which utilizes similarity measures.