as_proto_net: Convert Various Objects to 'proto_net's

In knapply/tweetio: Really, Really, Ridiculously Fast I/O, Transformation, and Analytical Routines Twitter Data

Description

A proto_net is a list containing two data frames named edges and nodes.

Usage

as_proto_net(
  tweet_df,
  target_class = c("user", "hashtag", "url", "media"),
  all_status_data = FALSE,
  all_user_data = FALSE,
  as_tibble = tweetio_as_tibble(),
  ...
)

## S3 method for class 'data.frame'
as_proto_net(
  tweet_df,
  target_class = c("user", "hashtag", "url", "media"),
  all_status_data = FALSE,
  all_user_data = FALSE,
  as_tibble = tweetio_as_tibble(),
  ...
)

## S3 method for class 'data.table'
as_proto_net(
  tweet_df,
  target_class = c("user", "hashtag", "url", "media"),
  all_status_data = FALSE,
  all_user_data = FALSE,
  as_tibble = tweetio_as_tibble(),
  ...
)

Arguments

tweet_df	A data frame of tweets, as obtained by read_tweets() or one of {rtweet}'s collection functions, e.g. rtweet::search_tweets().
target_class	character(1L), Default: "user". The class of nodes to use as the second half of each dyad (target/to/head). See Details.
all_status_data	logical(1L), Default: FALSE. Whether to attach all relevant status data to the edges data frame, which can then be used as edge attributes for downstream tasks.
all_user_data	logical(1L), Default: FALSE. Whether to attach all relevant user data to the nodes data frame, which can then be uses as node attributes for downstream tasks.
as_tibble	<logical>, Default: tweetio_as_tibble(). Whether a tibble::tibble() should be returned. Ignored if the {tibble} package is not installed.
...	Arguments passed to or from other methods.

Details

• In a proto_net, users are always to source/from/tail side of dyads. target_class defaults to "user", which creates edges where users on both sides of dyads.

  • However users can also share edges with "hashtag"s, "url"s, or "media", so those values are also valid to provide to target_class to create two-mode/bipartite proto_nets.

• The edges of a proto_net represent the statuses that form each tie, and status-specific columns are attached to the edges.

• Casing

  • Twitter hashtags are not case-sensitive, so if target_class is "hashtag" they will be cast to lower-case so they can represent the same node in downstream tasks.

  • URLs (after the domain) can be case-sensitive, so they are left as-is.

    • If you decide to cast URLs to lower-case yourself, exercise caution with Twitter's media URLs, as they contain a case-sensitive hash.

• all_status_data and all_user_data default to FALSE as they can be performance bottlenecks for large data sets, but they provide a way of building richly decorated networks with the maximum amount of attribute data embedded in the graph structure.

Author(s)

Brendan Knapp [e-mail]

Examples

path_to_tweet_file <- example_tweet_file()

tweet_df <- read_tweets(path_to_tweet_file)

tweet_df %>%
  as_proto_net(as_tibble = TRUE)

tweet_df %>% 
  as_proto_net(target_class = "hashtag", as_tibble = TRUE)
  
tweet_df %>% 
  as_proto_net(target_class = "url", as_tibble = TRUE)
  
tweet_df %>% 
  as_proto_net(target_class = "media", as_tibble = TRUE)
  
tweet_df %>% 
  as_proto_net(all_status_data = TRUE, all_user_data = TRUE, as_tibble = TRUE)

knapply/tweetio documentation built on Dec. 22, 2020, 7:15 p.m.