How do we calculate significant changes?

To calculate Significant Changes we look at each theme's recent performance to find those with volumes or scores more than 2 standard deviations from their recent average.

We then filter for statistical significance (p<0.05), and rank them by the magnitude of change from average.

There are 3 kinds of changes we can highlight.

1 Rare themes on the rise

These are smaller themes that have had a significant increase in volume. This is good for highlighting themes that might not get much attention due to their size but can indicate something new is happening in your data.

Example: The theme weather delays for the past 6 months has had an average volume of 0.96%, and ranged from 0.67% to 1.24%. This month however it's spiked to 1.8%, an increase of 0.84% from the average.

2 Volume changes

This highlights the biggest increases and decreases in theme volume, regardless of their original size. This is good for identifying themes that are having increases or decreases beyond expected fluctuation based on the last 6 months.

Example 1: The theme check in and boarding for the past 6 months has had an average volume of 25.46%, and ranged from 23.96% to 26.96%. This month however it's spiked to 28.71%, an increase of 3.24% from the average.

Example 2: The theme food and drinks for the past 6 months has had an average volume of 26.59%, and ranged from 24.49% to 28.7%. This month however it's dropped to 23.43%, a decrease of 3.17% from the average.

3 Score changes

This highlights the biggest increases and decreases in theme score. This is good for identifying themes that are having increases or decreases in their score (and possibly impacting your final score) beyond expected fluctuation based on the last 6 months

Example 1: The theme seat recline for the past 6 months has had an average NPS score of -55.39, and ranged from -69.36 to -41.42. This month however it's spiked to -7.96, an increase of 47.4 from the average.

Example 2: The theme economy comfort for the past 6 months has had an average NPS score of -7.17, and ranged from -33.07 to 17.72. This month however it's dropped to -57.14, a decrease of -49.97 from the average.