Over the years I collected some valuable data aphorisms. Let me share them. Some of them might save your career, or sanity.
“I know you said that this data was misleading, but it’s all we’ve got”.
Don’t search under the lamppost just because it is illuminated. Search where the key is likely to be!
First, do no harm. If you cannot give an answer based on data, say so! That way, you can make decisions with the necessary humility.
“It might not be a very accurate model, but at least it’s based on evidence, so it’s better than guessing”.
Based on evidence, man feed turkey out of kindness… until Thanksgiving!
“We cannot accept this outcome!”
Everything has a price, even your life, so prevent that outcome or be wrong and discredited.
“Data has meaning only in context.”
A number without a unit or a standard to be compared is just a number, and should be treated as such.
A number without a name or a history is just a number, and should be at best be ignored.
Ask yourself where the data came from, whether it is complete, who gave it to you and what interests they may have. Do you know what was computed, and what might have been left out?
“Insufficient data for [a] meaningful answer.”
— Multivac (Isaac Asimov)
If you can figure out how to reduce entropy in a closed system, please call me or the gods themselves.
This one, from Multivac, is my favorite data related quote. It answers the last question —what p-value to choose :).
“If you torture the data long enough, it will confess to everything.”
— Ronald Coase
“Does the data accurately represent the reality it abstracts from?”
The art of working with data consists mainly of understanding these last two quotes.
I hope you found these data aphorisms useful. Feel free to comment, so we can learn from each other.
