1.8.3 FAQ: Cluster Summaries & Focus Modal

Do the redesigned Cluster Summaries and Traits apply to all of my existing reports?

As of July 26, 2018, all reports display the redesigned Cluster Summaries and Traits. 

 

What has changed about Traits?

Now you can:

  • Find interests among other Traits
  • Select to view over- or under-indexing Traits for a given cluster
  • Filter Traits by metric types such as hashtags, bio keywords, and interests (available metrics vary based on report type)
  • Click-through traits to open the Focus Modal for additional context such as cluster penetration, audience comparisons, examples, and more

How is the average age calculated for a cluster?

The average age is calculated based on the median of each age group and the number of members that fall into each group.

What does the “audience” number represent?

The "audience" number represents the average age or gender demographic of the overall audience.

What do the color bars represent in the Overview of the Focus Modal?

The color bars represent how a metric is observed across all clusters within an audience, similar to the normalized bars you see throughout the Affinio dashboard.

1.8.3 Metric_Overview-1

How do I interpret Related Traits?

Related Traits can help you uncover cross-cluster behavioral patterns and gain a multidimensional view of metrics. Related Traits can be of any metric such as hashtags, links, bio keywords, and interests. Traits are considered related when we observe a similar "cluster footprint." You'll see the cluster footprint visualized throughout the Affinio dashboard on bar chart modules (i.e. normalized bars) and along the bottom of other metrics such as interests and links.

The distance from the center and darkness of the line are indicative of how similar the patterns are between traits. For example, here you’ll see friar_martin_dc is closest to #cmafest which means that they have the most similar pattern of cluster occurrences.

1.8.3 Related_Traits

You can explore more Related Traits by clicking through to another metric on the graph.