The Australian Twittersphere is a longitudinal collection of tweets from a population of Twitter accounts which explicitly self-identify as Australian or having connection to Australia. A publication detailing the method used to determine this population is forthcoming. This dataset contains the rules for inclusion in the Australian Twittersphere which are derived by the population method. These rules were developed by the QUT Digital Observatory team in 2021, and then used to determine the Twitter accounts which were included in the Australian Twittersphere up until the collection ceased in 2023.

To use these rules:

The rules file can be used with the twittersphere tool to filter a list of Twitter accounts down to those which match the rules and as such would have been suitable for inclusion in the Australian Twittersphere.

File specification:

aus_ts_rules_2021-09-16.csv is a CSV file encoded as UTF-8, with commas as field separators and each value enclosed in double quotation marks. There is a header row with column titles at the start, and each subsequent row (10,940 in total) represents a single rule in the ruleset.

Columns:

• include: 1, -1, or blank
  • 1 indicates that this is a positive matching rule
  • -1 indicates that this is a negative matching rule
  • blank: does not affect results, but this rule is retained to indicate that it has already been assessed
• discuss: Non-blank values indicate that this rule requires further discussion for finalisation
• note: Any notes
• field: Which field of the Twitter account this rule is to be applied to. Options: 'location', 'description', or 'realname'
• first_token: Token to match on to apply the rule
• second_token: Token to match on to apply the rule
• third_token: Token to match on to apply the rule

Note that first, second, and third tokens together represent a sequence of up to three tokens appearing in the specified field. If the relevant sequence for the rule is only one or two tokens in length, third_token and possibly second_token will be empty.