This document in OpenAlex has 340 authors. Many of them have 53 affiliations listed in the institutions array:
This disagrees with the actual document. For example, the lead author, Chekanov, has 53 affiliations according to OpenAlex, but in reality has only one:
I am finding a very large number of errors with author name disambiguation and affiliation matching across the entire data set. This is an addition to systematic errors with publication counts and derived values (e.g., h-indexes) that I have previously communicated with the OpenAlex team about. For example, in the most recent snapshot there is an author record with 434,622 attributed works:
Is the team aware of these sorts of problems, and if so, what are the plans to improve the data quality? My team's primary use case for OpenAlex is author and institution-level analytics, so these sorts of issues are major concerns.