Tumblr community analysis

tumblrAPI

#!/usr/bin/env Rscript


require(jsonlite)
require(reshape2)
api_key='your_api_key'



get_reblogs<-function(blogname,apikey,total=100) {
    res=lapply(seq(0,total-20,by=20), function(offset) {
        apiresults <- fromJSON(sprintf('http://api.tumblr.com/v2/blog/%s.tumblr.com/posts?api_key=%s&reblog_info=true&offset=%d&limit=%d',
            blogname,apikey,offset,ifelse(total-offset<20,total-offset,20))) 
        if('reblogged_root_name' %in% colnames(apiresults$response$posts)) {
            apiresults$response$posts[,c('blog_name','reblogged_root_name')]
        } else {
            data.frame(blog_name=NA, reblogged_root_name=NA)
        }
    })
    do.call(rbind, res)
}

reblogs = get_reblogs('glitchgifs',api_key)
for(blog in unique(reblogs$reblogged_root_name)) {
    reblogs = rbind(reblogs, get_reblogs(blog, api_key))
}

# filter reblogs in glitchgifs network
dataframe=reblogs[reblogs$reblogged_root_name %in% unique(reblogs$blog_name),]

# filter self reblogs
dataframe=dataframe[apply(dataframe, 1, function(x) all(!duplicated(x))),]

# write raw blog data (e.g. has duplicate edges)
write.csv(dataframe, file="reblogs.csv",quote=F)


# cast to matrix
m=acast(dataframe,reblogged_root_name~blog_name)

# write adjacency matrix to file
write.csv(m, file="reblog_matrix.csv", quote=F)

# write adjacency list using remelted matrix to get edge weights
write.csv(melt(m), file="weights.csv", quote=F)

Example of the "glitchgifs" community after plotting in cytoscape

Example of the "glitchtheory" community after plotting in cytoscape