Detect correlations with time shift

[allo-]

Finding correlations between trackers should allow for (small) time shifts.

Example:

• Tracker 1: Number of cups of coffee
• Tracker 2: Hours of sleep

Drinking a lot of coffee today cannot change the hours of sleep one logged in the morning, but it may affect the hours of sleep that one will log tomorrow. So the amount of coffee from one day should be correlated with the hours of sleep entered on the next day.

This may also be related to the issue, that "Tomorrow" should not start at 00:00.
When using a sliding time window, the log entries for the coffees and the log entry with the amount of sleep in the morning would be in the same 24 hour window.