I think the hubski idea is scale-able. Structurally it is the same as twitter.
Because of this I think Twitter problems are more likely than Reddit problems. Twitter has super twits but if you don't like them just don't follow them. I think you have a point that hubski could promote super-users and cults of personality @kleinbloo@ an mk have around 6 times as many followers as I do the thing is there is hardly a split as a 3rd of my followers also follow klein and half follow mk. The system is too porous to allow tightly controlled fiefdoms.
There is power to high follower counts but not really any control.
The analysis I would like to see (and really, who better than you to do it?) is the average number of shares it take for a post to propagate through, say, 90% or 95% of the community. My guess is not that many. I think its something you could easily do with the analyses that you've already conjured up.
The problem is data scraping is I had an excel file of the graph at a given date I could do it.
Indeed. As I've mentioned before, scraping data from Hubski would be awesome! However it's quite tough to do right now since there's no API and lots of the actual userdata is retrieved and viewed dynamically via javascript (not served on static pages). I've been thinking about trying to get some kind of basic data gathering tools together for Hubski. I really need to set aside a weekend and make it happen :)
the followers of the top 30 most followed users is around 1550
so about 1/4 of the total accounts. If I had a excel file of the entire graph I could do a distance and centrality study. from messing with the thing I think it is quit shallow but deeping.
between 4 and 6 shares.