The Metadata from a Single Social Media Post

A detailed look at the type of data that can be found in every social media post

A short Tweet or Instagram post may seem simple, but there is a lot of data hidden beneath the surface of every social post. That data, known as metadata, provides more context and information to every post, helping you understand trends, uncover insights, and more.

What is metadata anyway?

To put it simply, metadata is data about other data. Of course, what’s contained in the metadata completely depends on the type of of data you’re looking at. The term metadata was first used to describe the physical card catalogs in libraries before transitioning to being a term to describe digital data. Another common use of metadata is around digital photographs, where the metadata can describe the size, resolution, creation date, camera used, and much more for every image.

In the context of social media analytics, metadata is all of the data about a particular social media post. Metadata about a social post could include location, author, or time of the post. While some of that data can be seen in the social feed where that post appears, a lot of metadata is hidden behind the scenes. Metadata can be obtained from developer APIs or other raw data exports (often in the form of a .csv or other spreadsheet format).

In 2010, Raffi Krikorian (then a Twitter Engineer), illustrated the anatomy of a tweet featuring diverse metadata available through the Twitter API. The Twitter post provides a good example of the type of data that can be obtained from most social media posts across platforms.

How is metadata useful?

Social metadata is particularly useful in the analysis of social media data. It can help us organize, store, and archive data efficiently to query later. In the case of social media data, metadata allows historical posts to be analyzed effectively.
Metadata reveals new dimensions to filter through and explore when analyzing social media data, such as the time frame, audience, sentiment or location of a social media conversation. More robust metadata makes even deeper analysis possible. This allows brands to understand an audience, track brand health, uncover customer trends and more.

What kind of metadata can you find in a social post?

Below is an example of metadata from a single social media post as gathered in a Crimson Hexagon monitor. The example is in JSON format, a standard output format from an API call.
In this JSON example, metadata is organized by “keys” and “values”:

  • “keys”: analogous to column headers in a spreadsheet (in quotes and in red below)/li>
  • “values”: analogous to individual cells within rows in a spreadsheet (always follow the “key”)

Here’s a breakdown of the metadata categories in the example:

Location and language metadata

  • Location of the post (self-reported)
    • The location data of the post as selected by the author. For example, someone might post a photo from an event the day after. On Instragram, they may choose select the location of where the photo was taken instead of their location at the time of the post.
  • Expanded location details
    • These are the specific details of the location include country, state, city. Depending on the post, this may be as specific as the exact coordinates of the location.
  • Language detected within the post
    • Identification of the language in the post based on the text in the post.

Author metadata

  • Author of the post
    • The username of author of the post
  • Klout score of the author
    • Klout scores are a measurement of the reach and influence of a specific user
  • Author activity metrics (at time of post)
    • Activity metrics include additional info about the author (depending on the social channel of the post) including:
      • Number of posts
      • Number of followers
      • Number of accounts author is following
      • And more

Sentiment and emotion analysis metadata

  • Sentiment category name
    • A tag of Positive, Negative, or Neutral for the sentiment of the post.
  • Sentiment category probability
    • The raw numbers for the measurement of sentiment used to determine the sentiment category.
  • Emotion category name
    • The emotion (joy, anger, fear, sadness, disgust, surprise) identified in the post, if the post has an identifiable emotion
  • Emotion category probability
    • The raw numbers for the measurement of emotion used to determine the emotion category.

Image analysis metadata

  • Link to the image
    • A direct link to the image file, if there is an image attached to the post.
  • Name of detected image “class” (logo, object, scene or action)
    • If there are any identifiable logos, objects, scenes or actions within the image, they will be named here.
    • Example from the JSON code above:
      • Classes identified:
        • Flight
        • Macro photography
        • Vehicle
        • Helicopter
    • Image from the example:
  • Probability that the class (logo, object, scene or action) is present in the image
    • The raw numbers on the probability that the identification of a logo, object, scene or action is accurate.

With all of this detailed data from one individual social post, you can imagine the massive amount of data contained in over one trillion social media posts. Analyzing that massive data set can help uncover powerful insights about brands, products, consumers, or any other topic.

Request a Demo

Ready to transform your business?

Get a walkthrough of Crimson Hexagon and learn how consumer insights can help you make better business decisions.