Data qualitycomes up in most presentations I give to research audiences about the Atlas of Living Australia. The usual comment is How do I remove the amateur observations from the data in the Atlas? My response is that many amateur observations are of better quality than scientific ones and that fitness for use is a far better term than data quality because a record that one scientist filters out may be appropriate in other circumstances. It took me around a year to ensure that the Atlas of Living Australia made public all submitted records. My argument was that scientists wouldnt appreciate data hidden from them. Therefore, how do Data Publishers address this issue? I say that we can provide most value by the application of a standard suite of automated tests to all records, and where failed tests are flagged in the record. Such tests cannot detect and address all issues but they are easy to apply and are effective in determining fitness for use. If such tests are comprehensive, adequate for filtering purposes and representative in addressing known issues, then why not standardize them? A standard suite of tests makes life easier by consistency across Data Publishers.
