Learning path / How to detect signal from noise and build powerful filtering rules

Step 2: Identifying what “signal” means to you

 


This is step 2 of the learning path, How to detect signal from noise and build powerful filtering rules.



One of the most important steps towards building effective filtering rules and matching on Tweets of interest is to identify the “sources of signal” that matter the most to you. In other words, to build a strong ruleset and ensure that you receive Tweets of interest, you first need to: 

  1. Define exactly what information you’re trying to uncover with Twitter data; 
  2. Know exactly which Tweets you want to target with your rule(s). 
     

Work your way through the following prompts to identify exactly what signal you’re looking for. For each prompt, you’ll find examples of filtering operators that you can use.


Tweets from specific accounts

Are you interested in Tweets that are linked to specific accounts?

  • These could be Tweets that are sent from or to a given account; or Tweets that mention a specific user.
    Use from:, to:, and @
  • Accounts of interest might include leaders or influencers in a field, industry professionals, as well as official brand handles.
    Use is:verified, bio:, bio_name:statuses_count:, followers_count:, friends_count:, and listed_count:
     


Keywords and exact phrase filters

Use relevant keywords and phrases to target industry discussion. 

  • What words or phrases are typically used to refer to your area or object of interest?
    Use keyword, "exact phrase match", "keyword1 keyword2"~N, emoji, and contains:
  • Are any hashtags or links typically present in Tweets that matter to you?
    Use #, url:, url_title:, url_description:, and url_contains:
  • Are you looking for Tweets in specific languages only?
    Use lang:
     


Tweets with specific attributes

Are you looking for Tweets that contain specific attributes or “attachments”?

  • Tweets that have images, videos, or media:
    Use has:images, has:videos, and has:media
  • Tweets that have URLs, hashtags, or symbols:
    Use has:links, has:hashtags, and has:symbols
  • Tweets that have been classified by Twitter as being of a particular language (in other words, where the language value is not "undefined")
    Use has:lang
     


Tweets of a specific type

  • Are you looking for specific types of Tweets (Retweet, reply, or Quote Tweet)?
    Use is:retweet, is:reply, and is:quote
  • In order to match on original Tweets only, you can exclude all other types of Tweets, by using a hyphen to negate the following operators, like so:-is:retweet -is:reply -is:quote

Please note: A Retweet is a Tweet that has been “reposted” with no additional comment. A Quote Tweet is a Tweet that has been “reposted” with a comment.
 


Tweets with specific geo attributes

  • Are you interested in Tweets that were explicitly sent from a specific country or region?
    Use place:, place_country:, point_radius:, bounding_box:, profile_country:, profile_region:, profile_subregion:, profile_locality:

Please note: Most Tweets do not have geo attributes attached to them and using any of the available geo filtering operators will significantly narrow the scope of your request. 
 



A full list of available enterprise operators and associated descriptions can be found in our enterprise rules and filtering documentation

Please note:

The Search API only supports a subset of the filtering operators available with the PowerTrack API. 

 

 

 

 

 

Go to the next article - Step 3: Building initial filters

Go back to the learning path homepage - Learning path: How to detect signal from noise and build powerful filtering rules