Seriously though, it's interesting how interconnected some things can be in this view. I'm not sure what sense I can make of those interconnections, though. Mousing around, while being a very frequent redditor (so my own neural network is making connections based on experience), I can kinda infer order out of things like the "government->state" topics connected to "force" and "property" among others (hints at the libertarian-leaning general population), and the "women" topic connecting to a whole host of stuff...the cyan colored section off to the top right might even kinda hint at the casual misogyny thing (which was a "ha ha only serious" kind of joke), with words like "bullshit", "logic", "proof", "assumption", "reasoning", and "evidence", being connected to "women" but not to "men".
But, without having spent years on reddit, and without my particular flavor of reddit (the subs I'm subscribed to), maybe I'd interpret the data very differently. I never quite know how to interpret network graphs like this, honestly, short of for things that are networks. i.e. a computer network topology on a graph shows useful data...the hops from one machine to the next. When connecting up one word to the next, it seems difficult to draw meaningful conclusions. Like my interpretation of the meaning of "government->state->property" as being a hint at the libertarian leanings of many subreddits, or the connection of "women->reasoning->evidence" as being a hint of many redditors belief that women are illogical liars (which is the impression many of my female friends have of reddit, in general, particularly when topics like date rape or the "friend zone" come up). Is that actually the context in which these connections are made? I wouldn't really know how to check. It'd be cool to be able to drill down to conversations in which the connections where made, but presenting that in a coherent UI seems challenging.