I am trying to figure out good model to analyze social dynamics using a model. I have commenting data from a website. Such as 500 articles, 10000 users, and their 75000 comments. Every user does not comment on every article.
The opinions posed by each user is linked to the opinion of the previous users, as well as the opinion of the article itself. The opinions of previous users are also linked to the previous users and the article.
I want to estimate the impact of opinion of previous user, average user opinion, article opinion and variation of opinions on the current opinion. Can you suggest a methodology for such a problem? Everything is correlated.
I also would like to ask, what software and code would you use to model a problem like this? Codes are very very appreciated.