Scala: Window.partitionBy() and lag function to determine largest % increment from previous day

Question

I have data that is being read in from two separate .csv's with the following headers:

df1:

ID1 (integer), ID2 (integer), person_count (integer), distance (float), amount (float), payment_type (integer), datetime_origin (string), datetime_dest (string)

df2:

ID1 (integer), city (string), zone (string), zone_service (string)

I need to find the 3 days of the month April where city = "LA", that saw the largest percentage increment in pickups compared to previous day.

Any ideas on where to begin solving this? I think I might need a lag function over a Window.partitionBy() for the day of the month, but not sure as far as implementation goes.

Thanks!