Skip to content

Methodology

Index Scoring

These scores are an index where 100 is the national average. An index of 205 for the Fitness Enthusiasts segment means that the devices in this particular block group are 2.05x more likely to have visited fitness enthusiast related content online than the national average. How is that calculated? The full formula:

\[ Index \, Score = \frac{Segment \, Percentage}{National \, Percentage} * 100 \]

Where:

\[ \frac{Segment \, Percentage}{National \, Percentage} = \frac{N \, Segment \, Devices \, in \, BG / N \, Devices \, in \, BG}{N \, Segment \, Devices \, in \, USA / N \, Devices \, in \, USA} \]

Missing Values

There 2 kinds of "missing" values:

  • Not enough data in a given block group
  • Enough data in a given block group, but not enough for the given segment

When there is not enough data in a given block group, no values for any segment are provided.

When there is not enough data in a given block group for a certain segment, but there is enough data to provide scores for other segments, then that segment receives an index score of zero. The threshold below which a segment will receive a zero score is variable according to the base rate of the segment. As a result, less common segments (smaller audiences) are more likely to have zero values.

Data Collection

Mobile IDs are used to bring desktop behaviors to the physical world. Using a probabilistic approach, mobile IDs are connected to a users' devices and common evening location from mobile movement data determines a users' home location. Scores for a given block group represent the aggregation of device behavior from the devices whose home location could be determined with a high degree of certainty.

Source

This dataset is provided at a geographic level in partnership between Spatial.ai and leading customer audience solution company Dstillery.


Back to top