Pick a city for its full deletion profile — or explore the country-wide picture below.
About the methodology and data sources
01Background
Recently, Google started allowing businesses in Germany to remove reviews by citing defamation laws. This annoys me, just like many other people. I saw people from smaller cities — because there are fewer venues there and it's easier to collect statistics — doing similar analyses with Excel spreadsheets. So I decided to build an interactive map.
02Data Collection
The main challenge is collecting information from Google Maps.
I did not collect the reviews themselves, only statistical data — specifically the current rating, the number of reviews, and the number of deleted reviews. Google recently started showing approximately how many reviews were removed. For example, Spaccanapoli Nr. 12 removed between 201 and 250 reviews:
201 bis 250 Bewertungen aufgrund von Beschwerden wegen Diffamierung entfernt.
03Methodology
Based on this data, I calculated the adjusted score of a venue — as if the reviews had never been removed.
Formula
adj_score =
(current_number_reviews × current_rating + deleted_reviews_estimate × avg_deleted_rating)
/ (current_number_reviews + deleted_reviews_estimate)
deleted_reviews_estimate = (deleted_reviews_max + deleted_reviews_min) / 2
avg_deleted_rating = 1
avg_deleted_rating = 1 is the approximate average rating of deleted reviews. In other words, I assume that every deleted review was, on average, a 1-star review. That's probably not entirely accurate, but it makes sense to calculate the worst-case scenario.
Example — Spaccanapoli Nr. 12
current_rating = 4.8
current_number_reviews = 1266
deleted_reviews_estimate = (250 + 201) / 2 = 225.5
adj_score = (1266 × 4.8 + 225.5 × 1) / (1266 + 225.5) = 4.2
Every trust score is computed from a venue's public reviews using a published method: what is collected, how the score is built, what the “unusual patterns” figure counts, and where the approach is weak. Review text and author details are deleted once the analysis is computed.
Social media
My Projects
GPTransport
Real-time public transport departures for Germany