One mistake that I see quite often in the Discogs data is that people use the "Barcode" field for basically everything in the section "Barcode and Other Identifiers" (or "BaOI" in Discogs lingo). This is not very surprising, as "Barcode" is the default value and no checking is done. Almost two years ago I already
checked in how many releases the "Barcode" field was used to store the value of a rights society and found around 400 releases, which resulted in the following graph:
|
Rights societies in the "Barcode" field in January 2018 |
although I honestly cannot remember how robust my checks were at that time (I believe they missed quite some data). I thought it would be interesting to see what I would find in the current data set. The result: 1118 unique releases, and distributed across the data as follows:
|
Rights societies in the "Barcode" field in October 2019 |
So that is significantly more. A few things that are noteworthy: there are quite a few more for each month (meaning that I indeed catch more mistakes than two years ago), but there is also a very significant spike somewhere in early 2018. I have no idea what happened there. I guess we're back to cleaning those up...
Comments
Post a Comment