{"id":12510,"date":"2019-09-14T09:11:23","date_gmt":"2019-09-14T14:11:23","guid":{"rendered":"https:\/\/rud.is\/b\/?p=12510"},"modified":"2019-09-17T06:40:57","modified_gmt":"2019-09-17T11:40:57","slug":"twitter-account-analysis-in-r","status":"publish","type":"post","link":"https:\/\/rud.is\/b\/2019\/09\/14\/twitter-account-analysis-in-r\/","title":{"rendered":"Twitter “Account Analysis” in R"},"content":{"rendered":"

This past week @propublica linked to a really spiffy resource for getting an overview of a Twitter user’s profile and activity called accountanalysis<\/a>. It has a beautiful interface that works as well on mobile as it does in a real browser. It also is fully interactive and supports cross-filtering (zoom in on the timeline and the other graphs change). It’s especially great if you’re not a coder, but if you are, @kearneymw’s {rtweet}<\/a> can get you all this info and more, putting the power of R behind data frames full of tweet inanity.<\/p>\n

While we covered quite a bit of {rtweet} ground in the 21 Recipes<\/a> book, summarizing an account to the degree that accountanalysis does is not in there. To rectify this oversight, I threw together a static clone of accountanalysis that can make standalone HTML reports like this one<\/a>.\"twitter<\/a><\/p>\n

It’s a fully parameterized R markdown document, meaning you can run it as just a function call (or change the parameter and knit it by hand):<\/p>\n

rmarkdown::render(\n  input = \"account-analysis.Rmd\", \n  params = list(\n    username = \"propublica\"\n  ), \n  output_file = \"~\/Documents\/propublica-analysis.html\"\n)\n<\/code><\/pre>\n
It will also, by default, save a date-stamped copy of the user info and retrieved timeline into the directory you generate the report from (add a prefix path to the save portion in the Rmd to store it in a better place).<\/p>\n
With all the data available, you can dig in and extract all the information you want\/need.<\/p>\n
FIN<\/h3>\n
You can get the Rmd at your favorite social coding service:<\/p>\n