Visualise Twitter Interactions

Allows building an edge table from data frame of tweets, also provides function to build nodes and another create a temporal graph.

Build Status Build status Coverage Status CRAN_Status_Badge CRAN log twinetverse

Visualise networks of Twitter interactions.


install.packages("graphTweets") # CRAN release
devtools::install_github("JohnCoene/graphTweets") # dev version


  • gt_edges & gt_edges_bind - get edges.
  • gt_co_edges & gt_co_edges_bind - get co-mentions
  • gt_nodes - get nodes, with or without metadata.
  • gt_dyn - create dynamic graph.
  • gt_graph - create igraph graph object.
  • gt_save - save the graph to file
  • gt_collect - collect nodes and edges.

See the documentation for guides and examples.


graphTweets 0.5.1

Better use of tidyeval to improve usage of the package in functions.

graphTweets 0.5.0

  • gt_co_edges replaces gt_edges_hashes, it now works with other columns than just hashtags.
  • gt_edges_hash is deprecated in favour of gt_edges.
  • gt_edges_bind and gt_co_edges_bind added to bind edges together and build more complex graphs.
  • Escape hatches, functions ending in _ are no longer necessary and are thus deprecated.

The origin gt_edges functions, first part of the package in 2014, was extracting @tagged users from tweets' text with convoluted regular expressions: it is no longer necessary as rtweet now returns mentions_screen_name, hence the aforementioned changes.

graphTweets 0.4.3

  • Added gt_edges_hashes_ and gt_edges_hashes to build networks of #hashtags co-mentions.
  • Added %<-% from the zeallot package to unpack the nodes and edges.
  • Fixed gt_dyn bug where lifetime was not working properly.

graphTweets 0.4.2

Removed splitstackshape dependency ahead of its archival; now uses tidyr.

graphTweets 0.4.1

  • gt_nodes returns number of n_edges, the number of edges the node is present in.
  • gt_edges_hash and respective escape hatch added
  • gt_dyn returns correct start and end.

graphTweets 0.4.0

Major release: overhaul to 1) make computations much faster, 2) make the whole package more tidyverse friendly and 3) switch to rtweet as main source.

  • getEdges & getNodes are now deprecated in favour of gt_edges and gt_nodes
  • dynamise deprecated in favour of gt_dyn
  • magrittr pipe added.
  • gt_collect added: use to get to collect edges and nodes as list.
  • gt_graph added: use to convert to igraph object.


token <- create_token("APP", "xxxXXxxxx", "xXXXxxXX")
tweets <- search_tweets("#rstats", token = token)
  "v3.2" = {
    edges <- getEdges(, "screen_name", "text")
    g <-, TRUE)
  "v4" = {
    tweets %>% 
      gt_edges_() %>%  
      gt_graph() -> g
  test replications elapsed relative user.self sys.self user.child sys.child
1 v3.2          100    6.55    1.492      6.45     0.06         NA        NA
2   v4          100    4.39    1.000      4.33     0.05         NA        NA

graphTweets 0.3.2

  • documentation corrected
  • Better extract @handles
  • Better cleans @handles (i.e.: white space)

graphTweets 0.3

  • Fixed #1 issue, input check in dynamise
  • Added dynamise

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.5.3 by John Coene, 2 years ago

Report a bug at

Browse source code at

Authors: John Coene [aut, cre]

Documentation:   PDF Manual  

Task views: Web Technologies and Services

MIT + file LICENSE license

Imports dplyr, igraph, purrr, rlang, magrittr, utils, tidyr, zeallot, combinat

Suggests rtweet, testthat

See at CRAN