How is Cross App Usage Calculated?

Apptopia's Cross App Usage uses Google Play or Apple App Store review data to determine what other apps are likely used most by a given app's user base

Apptopia's Cross App Usage data is an index which will help you better understand which other apps are most commonly used by a particular app's user base.  For instance, you could learn that active users of Tinder are also actively using Bumble and Hinge.  You could even deduce that the overlap between Tinder and Bumble was 2x more likely than the overlap in users between Tinder and Hinge.
 
Here is a look at the tool using Draft Kings Fantasy Sports as an example:
Apptopia cross app usage
 
The primary data set we use to build our index is 1.5 billion user reviews.  Each review is tied to a unique store user ID and we can see all the apps a user reviews. Using store UserIDs, we can build a detailed matrix which measures the amount of audience overlap between 2 apps (when a user reviews two or more apps).
 
Visualization bars were created so you can easily see the amount of overlap each has with a given app's user base.
Just to the right or the visualization bars is the % change in weight by time period set in the date drop-down. 
 
app overlap weight
If you mouse over the bars, we showcase an "Overlap Weight", this data point is calculated by dividing the [Total Overlaps between App A and App B] / [ Total Possible Overlaps...i.e. total reviews, of App B].  Using App B as a denominator is important as this is the reason you don't see Facebook & Google Maps at the top of every Cross App Usage chart.  Instead the above Weight highlights the strongest cross app usage based on size and relevancy.
overlap weight example
 
cross app usage example
 
The Overlap Weight is not meant to tell you the exact # of users who use both apps, but rather is meant to be used to understand proportions.  For instance If App A and B had an Overlap Weight of .80 and App A and App C had an overlap weight of .2, you could confidently assume that users of App A are 4x more likely to also be active users of App B versus App C.