Network Of Mentions • ConnectR

This article plays through an exemplar workflow for creating a network of mentions.

This can be used to try to find communities of authors who all mention one another.

We’ll use the example data in the ConnectR package.

Will filter it a little because mentions are numerous.
Will also tidy up the names of the variables we want to use.

example <- ConnectR::mention_example %>%
  dplyr::sample_n(100) %>%
  dplyr::rename(followers = `twitter followers`, reach = `estimated reach`)

Create the network

Using a variable

Radarly exports provide mentioned authors in a variable called mentions and separated by “;”

To use this variable, set process = "radarly" and to = mentions.

radarly_mentions <- example %>%
  ConnectR::make_mention_network(process = "radarly",
                                 from = author,
                                 to = mentions,
                                 size = followers,
                                 colour = reach)

Extracting the mentions directly

If the dataset doesn’t contain a mentions variable, then ConnectR can be used to try to extract mentions from the text itself.

To do this, set process = "extract" and provide the name of the text column in your data as the to = argument.

extracted_mentions <- example %>%
  ConnectR::make_mention_network(process = "extract",
                                 from = author,
                                 to = text,
                                 size = followers,
                                 colour = reach)

Visualise the network

Radarly

Interactive

The default visualisation is interactive. You can:

Hover over a node to highlight its connections, name, and follower count.
Hover over an arrow to see how many retweets it represents.
Zoom in or out to see all the names for a collection of nodes.
Use the “edit” button to remove irrelevant nodes and edges.
Double click on a node to collapse all its edges.
Click and drag the nodes to make the layout clearer.
Search for a particular author in the dropdown menu.

ConnectR::viz_network(tbl_graph = radarly_mentions)

Static

It’s also possible to create a static visualisation of the network in case this is necessary.

ConnectR::viz_network(tbl_graph = radarly_mentions,
                      type = "static",
                      label_prop = 0.2)

Extracted

ConnectR::viz_network(tbl_graph = extracted_mentions)