Would it be possible to qualify the specific actions that would help growth in usage? I.e. correlating this dataset with specific "marketing" patterns such as use of push notifications, share buttons, etc.<p>How much does the app quality (ratings) correlate with usage patterns? If we assume that most apps aren't very high quality or don't bring much value to user, these findings would be kind of obvious; users find out that app isn't worth it and stop using it.<p>Overall, it would be interesting to know which apps break this pattern and why.