0
$\begingroup$

I am compiling a list of all of our customers' revenue, which will be stored as an integer in our analytics app (ChartMogul).

This revenue data is from Companies House in the UK. With small businesses (usually under £10m in revenue), they don't need to report their revenue, and therefore have an 'Unreported' value.

So 'Unreported' can mean anything from £0-9.99m in revenue, a big range.

What would be a sensible integer value I can use for these 'Unreported' businesses?

Something like -9999 which would allow us to report on those Unreported businesses, without making assumptions about what their actual revenue is?

What is best practice here?

Improve this question
$\endgroup$

1 Answer 1

Reset to default
0
$\begingroup$

In this case, since you are not really analyzing but reporting/storing the data, a NULL type or value that clearly is wrong (like -9999) would be common to denote missingness.

However, I would make sure to communicate the specific reason for the missingness that you highlight because any accurate analysis should take into account that the data is censored rather than missing. Knowing that the observation must be less than 10M is very useful and would need to be incorporated into any analysis to avoid biased results.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.