Any figure that looks interesting is probably wrong.
Twyman's law is the analyst's discipline. A striking number is far more often an error in collection, definition or joins than a genuine discovery, so surprise should trigger checking before publishing.
In practice
- A dashboard showing signups up 400% overnight is a duplicated feed four times out of five.
- An analyst who publishes a remarkable figure without checking the pipeline has usually published a bug.
- Further reading
- Wikipedia
- Subject
- Science & Epistemology
- Essay
- Paradoxes in the data, and sayings about knowing