Solved

Google Play Store's algorithm for ratings

  • 5 October 2023
  • 2 replies
  • 405 views

Badge

Hi to everyone,

I used data.ai API (specifically portfolio/app-store end) to obtain data about all the ratings of an app for the last month but I can’t figure out how to calculate the rating for that period from raw data like this:

I read online that Google is now weighing recent ratings more heavily than historical ratings, so using a simple formula like the following one will return incorrect rating:

RATING NUM = MAX(0,[@[one_star_incremental]])+2*MAX(0,[@[two_star_incremental]])+3*MAX(0,[@[three_star_incremental]])+4*MAX([@[four_star_incremental]],0)+5*MAX([@[five_star_incremental]],0)

RATING DEN=MAX(0,[@[one_star_incremental]])+MAX(0,[@[two_star_incremental]])+MAX(0,[@[three_star_incremental]])+MAX([@[four_star_incremental]],0)+MAX([@[five_star_incremental]],0)

OVERALL RATING FOR LAST WEEK = sum(RATING NUM) / sum(RATING DEN)

 

Can someone help figuring out what weights should I give to the most recent reviews?

Thank you!

icon

Best answer by rmartinshort 11 October 2023, 01:44

View original

2 replies

This is a good question and it’s about how to aggregate ratings for a given period  from incremental ratings provided by data.ai.

As far as I know, Google doesn’t share is method for “recency weightings” on the play store ratings. So there is no way to exactly replicate those ratings using the data we provide.

However, I still think your aggregation method is valid and can be used to compare apps if used consistently. Are you seeing major discrepancies ratings calculated in this way and what’s listed on the Play Store? 

Badge

Thank you Martin! yes unfortunately there’s a discrepancy and moreover the result doesn’t match with average rating displayed on data.ai dashboard for the same period ☹️ 

I replied with more details on this topic:

How can i calculate avg ratings from incremental avg ratings returned by API? | Community (data.ai)

Edoardo

 

Reply