The regions data in the languages was corrupted when the data set from the old Noto site was pulled in. Several steps were taken to return to a healthy state.
1. Clear regions for historical languages. These languages aren't used anywhere today.
2. Clear regions for languages with >10 regions. These languages are suspicious and will be repopulated at a later step if the data was correct.
3. Populate region data using CLDR.
4. Populate region data using the Omniglot merged YAML file from Adam with territory info.
