JPDB Kanji Data

Beyond 4000 there may be some missed that are classified under JPDB's Very rare or Never used categories but everything from all other categories is included and it seems likely to include the entirety of the Very rare category.

JPDB doesn't provide this data with exact numbers only in groups of 100. The Uncommon rating contains ranks 3600-4000. The Very rare rating includes anything past rank 4000 that is still within their corpus of which 1112 were found when scraping.

Data such as JPDB's keywords, readings, composition, examples, etc are also included.

This data has also been added into Kanji Grid under the JPDBv2 Kanji Frequency List grouping.

Frequency data already parsed is here https://github.com/Kuuuube/kanjigrid/blob/master/data/jpdbv2_kanji_frequency_list.json. If you want to parse the raw files for frequency yourself for some reason: (?).*?(?=)