2
$\begingroup$

We have build a POS system and we have an integration with some beer and soda dispensing machines. Currently we give the customer a difference between the total expected amount of liters used on a day and the actual liters used. However we actually recieve quite precise measurements from the dispenser and would like to be able to tell the customer which transaction doesnt match a dispense or the other way around.

However there is a difference in the time of the actual dispense and the time they put it into the POS (this time is usually a maximum of 15 minutes). Additionally the measurements show that the user often puts a different amount of beer into the glasses than they actually should.

To show you an example:

POSSystem
12:30 - 1 large beer (0,75L)
12:37 - 1 small beer (0,3L)
12:37 - 1 small beer (0,3L)
12:39 - 1 small beer (0,3L) and two large beers (0,75)
13:12 - 1 small beer (0.3L)
16:02 - 1 small beer (0.3L)

Dispenser log
12:33 - 0.5L beer
12:34 - 0.27L beer
12:40 - 0.62L beer
12:44 - 1.77L beer
15:55 - 0.3L beer

Now my question is, is there any way to match the dispenser log mathmatically to the best matches from the pos system. if i would look at these numbers by hand, i would guess that the beer sold at 13:12 was not actually poured.

As i dont have any clue how to solve this i have tried to add a tag that i thought made sense.

Hoping someone has an idea on how to solve this.

  • 0
    This is a really good question. I do not have an answer for you. You may find your answer on the [Cross-Validated SE](http://stats.stackexchange.com/). They deal with data analysis questions.2017-01-08

0 Answers 0