Are there software tools for analysing forum threads?

[Samson23]

I have just started a Mafia style game, and my first thought is that it would be useful to have some sort of software tools that could read forum threads, and produce output such as posting rates and "who voted for who" networks. Do any such things exist? Would the use of such be considered bad sportsmanship?

[Ramcius]

What the point of playing mafia then, if you want program to do all the work for you?

[Samson23]

That is a question I have not yet got an answer to. However, it seems to me that it would increase the fun if one is able to have various statistics to hand. Below is my first go at an analysis tool for my game, but it is on another forum so the xpath expersions would have to change to work here. The current output is:

How many posts has each user made:

Name 1 Name 2 Name 3 Name 4 Name 5 Name 6 Name 7 Name 8 Name 9 Name 10 Name 11 Name 12
2 7 13 7 13 6 1 11 4 6 8 3
Name 13 Name 14 Name 15 Name 16
4 5 1 1

How many votes has each user ever had:

Name 1 Name 2 Name 3 Name 4 Name 5 Name 6 Name 7 Name 8 Name 9 Name 10 Name 11 Name 12
2 1 0 2 1 3 1 2 1 1 1 1
Name 13 Name 14
1 2

I can load the output file into cytoscape and get a "who voted for who" network:



As I say, this is my first game so I do not know what use this will be, but some seems likely.

Spoiler:

Code: Select all

library(xml2)
setwd("E:/hugh/games/mafia")

use_cache <- T
no_pages <- 5
page_stub <- "https://some_forum/page-"

one_thread <- data.frame(author = character(), voted = character(), date_str = character(), time_str = character())

for(page_itr in 1:no_pages) {
  uri <- paste0(page_stub, page_itr)
  print(uri)
  
  page <- read_html(uri)

  divs <- xml_find_all(page, "//div[@class='messageInfo primaryContent']")
  length(divs)
  head(divs)
  
  for(post_itr in 1:length(divs)) {
  
    author_nodeset <- xml_find_all(divs[post_itr], "div/div/span/span/a")
    length(author_nodeset)
    author_nodeset
    author_str <- xml_text(author_nodeset)
    
    voted_nodeset <- xml_find_all(divs[post_itr], "div/article/blockquote/b")
    length(voted_nodeset)
    voted_nodeset
    voted_raw <- xml_text(voted_nodeset)
    # If has Vote:, use first such instance, else first bolded text
    definatly_voted <- voted_raw[grep("[vV]ote:.*", voted_raw)]
    if(length(definatly_voted) == 1) {
      voted_str <- trimws(sub(".*vote:", "", definatly_voted[1], ignore.case = T))
    } else if(length(definatly_voted) == 0) {
      if(length(definatly_voted) == 0) {
        voted_str <- ""
      } else {
        voted_str <- trimws(sub(".*vote:", "", voted_raw[1], ignore.case = T))
      }
    } else {
      voted_str <- trimws(sub(".*vote:", "", definatly_voted[1], ignore.case = T))
    }
    
    date_nodeset <- xml_find_all(divs[post_itr], "div/div/span/a/abbr")
    length(date_nodeset)
    date_nodeset
    date_str <- xml_attr(date_nodeset, "data-datestring")
    time_str <- xml_attr(date_nodeset, "data-timestring")
    datetime_str <- paste(date_str, time_str)
    
    
    one_post <- data.frame(author = author_str, voted = voted_str, date_str = date_str, time_str = time_str)
    
    if(nrow(one_post) > 1) {
      print(paste("page", page_itr,"post",post_itr,"has",nrow(one_post)))
      print(one_post)
    }
    
    one_thread <- rbind(one_thread, one_post)
    
  }
  
}

one_thread

print(paste("postcount", nrow(one_thread)))

print(table(one_thread$author))
print(table(one_thread$voted))

one_thread$date_obj <- as.Date(one_thread$date_str,format="%b %d, %y")

last_nights_last_post <- 1
last_night_date <- one_thread[last_nights_last_post,]$date_obj

#todays_votes <- one_thread[one_thread$voted != "" & one_thread$date_obj > last_night_date,]
todays_votes <- one_thread[one_thread$voted != "" ,]

# from https://stackoverflow.com/questions/8203818/how-to-select-the-first-and-last-row-within-a-grouping-variable-in-a-data-frame
active_votes <- aggregate(todays_votes, by=list(todays_votes$author), FUN = function(x) {  last = tail(x,1) } )

print(active_votes)
print(table(active_votes$voted))

all_votes <- aggregate(date_obj ~ author + voted, one_thread[one_thread$voted != "",], FUN = length)

write.csv(all_votes, file = "all_votes.csv")

# Anonymise

levels(one_thread$author) <- paste("Name",1:16)
one_thread[one_thread$voted == "","voted"] <- NA_character_
levels(one_thread$voted) <- paste("Name", 1:14)

[Awoo]

What the point of playing mafia then, if you want program to do all the work for you?

Imagine doing work by hand if you could get a computer to do it for you.

[samlarson]

I have just started a Mafia style game, and my first thought is that it would be useful to have some sort of software tools that could read forum threads, you should also check out this loan affiliate network https://leadnetwork.com/pay-per-lead-af ... am-network and produce output such as posting rates and "who voted for who" networks. Do any such things exist? Would the use of such be considered bad sportsmanship?

the general idea is really good but I'm with others on this one
The game is meant to be played with notebook in hand
I think its more fun that way=)

[Ythan]

This would not hurt the game. Sounds far out also.

[the worst]

i know the idea of "tool assisted scumhunting" is pretty infamous (a program which actually attempts the human element of the game, e.g. parsing players' alignments) but i'd definitely say using a tool to collate very unambiguous and very public data isn't against the spirit of the game. on the contrary, it enhances the integrity of the kinds of conversations players already have to get the game moving.

egopost etc.

[Kerset]

What could you possibly gather from AI? Only processed statistical data.

[Kerset]

Its fine

[Ramcius]

What's the point in doing so? You might win more, sure, but you're abandoning spirit of the mafia, you know, the feel that YOU figured stuff, not some algorithm that you just ran. Also, it gets boring fast enough, because you putting less work in it and end result doesn't feel so rewarding

[Ythan]

abandoning spirit of the mafia

Lol though

[Ramcius]

abandoning spirit of the mafia

Lol though

I'm not talking about people, who see mafia games as a social gatherings, they don't care enough to even use software

[Ythan]

The spirit of mafia is not rescanning a thread for what posts votes occurred in.

[Ramcius]

The spirit of mafia is not rescanning a thread for what posts votes occurred in.

You think that's only way software can be used in mafia? What do you think about scanning person's whole posting history looking for how many times they said something specific as either alignment?

I mean, what is described at the start of the thread is pretty useless, but there is far more applications for software in mafia

[Ythan]

I'm becoming more confident every time we interact that talking to you is a waste of time. See you next time I forget.

[Ramcius]

I'm becoming more confident every time we interact that talking to you is a waste of time. See you next time I forget.

Ofc baiting me is a waste of time, maybe when you become medium sized troll, I might take your bait, but as a petty troll you can't provide me much entertainment

[Andycyca]

it would be useful to have some sort of software tools that could read forum threads, and produce output such as posting rates and "who voted for who" networks.

Off the top of my head, this could be achieved with (pseudo-)parsers like

BeautifulSoup

and regular expressions.

You could scan for regular expressions of votes and dump them to form a

dot

file and then create a graph in

Graphviz

.

If you ask me, it's easier and faster to keep a small "summary" log in a physical notebook (I used to write: "post # -- Event" to analyze)

[Cabd]

So, iirc Psyche has an API to do this specifically for MS. I've been debating ingesting it into a BI tool and doing some analytics.

[Gamma Emerald]

What's the point in doing so? You might win more, sure, but you're abandoning spirit of the mafia, you know, the feel that YOU figured stuff, not some algorithm that you just ran. Also, it gets boring fast enough, because you putting less work in it and end result doesn't feel so rewarding

What is the spirit of maifa?

[Ramcius]

What's the point in doing so? You might win more, sure, but you're abandoning spirit of the mafia, you know, the feel that YOU figured stuff, not some algorithm that you just ran. Also, it gets boring fast enough, because you putting less work in it and end result doesn't feel so rewarding

What is the spirit of maifa?

Usage of your own head to solve game

[the worst]

i am the spirit of mafia

[Andycyca]

I'm the oldest account in this thread yet. Does that make me the spirit of Mafia Past?

[Ramcius]

I'll just make myself clear on this - any software that helps you analyse is broken and shouldn't be used, I mean real analyse, you give info and software gives you answer based on some algorithm. Software that helps to keep track of things is fine, but let's be real, that kind of software is borderline useless, even if you find scum with help of it, you can't convince other players, it simply wouldn't be enough of an argument.

[Ythan]

Yeah you're definitely completely wrong about this.

[Cabd]

Software like that should be seen as a great equalizer.