Songs: Eyeball The Data

notes / gigs / releases

We're gonna eyeball the data
And see what we've got
See what's there that shouldn't be there
And what should be there that is not

Are there weird repetitions
Does it suddenly change?
Check those null values and those zeroes
Cos those two fellsa are not the same

We're gonna eyeball the data
Look for columns that have got transposed
Are ANSI characters and UTF-8
Together? Should they be? No.

Check for numbers that are out of range
For instance in a weight or a height
Trailing spaces and carriage returns
Hiding in a text field's just not right

We're gonna eyeball the data
Until it's totally clean
So every variable has a description
Every table it's own UID

It won't be fun
And it's gonna take time
But when we're finished we'll have a dataset
On which we can rely

Eyeball the data
Let's get it done
When you've eyeballed the data
It can be shared with anyone

Published by Wipe Out Music Publishing

This song was originally written for the 'Data and Doctor Doom' show but then was taken out in very early drafts, before the whole thing was even finished, as it slowed down the pace and could very easily be cut in favour of a line saying "And then I cleaned the data".

It's a shame, because it's a really nice tune (according to me anyway), and it's based on some excellent advice that my old boss Professor Edward Melhuish always gave whenever we were doing any sort of data analysis. "Eyeball the data!" he'd say, meaning have a look at it to see if anything leaps out as odd. This was excellent advice which I still follow, and think about, to this very day, so I thought it deserved to be enshrined in song. Maybe I'll save it for when the show (inevitably) has its Broadway run and needs extending a bit!