Specimen · catalogued monster

The Data Swamp Thing

The data lake nobody dares wade into

Grown from years of tables tipped in and never fished out, it lurks beneath a thick green film of forgotten copies. Now and then it surfaces with something useful in its jaws, though nobody can say where it came from.

Symptoms

• No catalogue or metadata
• Duplicate, stale and cryptically named tables
• Unknown quality and ownership
• Analysts bypass the lake and go to source
• Data lands; nothing useful leaves

Overheard in the corridor

• “Is there a table for that somewhere?”
• “Nobody knows what's in that bucket.”
• “I'd rather pull a fresh extract than trust the lake.”
• “We've got three copies and none of them match the source.”