With apologies to Borges, I believe that all data sets can be classified as follows:
- Those that belong to the emperor.
- Those that are irrelevant.
- Those that are presently empty.
- Those that describe tulips.
- Those that are included in this classification.
- Those that make analysts tremble as if they were mad.
- Those originally written on parchment with a fine camel hair brush.
- Screenshots of spreadsheets embedded in PDFs.
- Those containing the word “NULL”.
- Those that include the only copy of the API key needed to access them.
- Those that when plotted at low resolution appear to be distributed normally.
- Those with indisputable provenance.
- Those whose columns are of monotonically increasing width.
- This one.