Generating additional metadata using the data itself

[jhp7e]

Existing metadata often omits valuable information that would be useful for researchers. For example, it may state that there is a variable called "station" but won't tell you how many different stations there are. It may tell you the date range, but not the frequency of sampling or the number of unique dates or it may have a variable called "species" but not enumerate how many distinct species there are. Generic code that would analyze the content of an arbitrary data table and produce a metadata report would solve this issue.