6  Standardisation

We might initially work with a dataset that hasn’t been standardised.

Why might this be a problem?

Patterns may be an artefact of population

Areas with more people may just have higher values - but it may not be unusual given the number of people.

Be careful with your interpretation!

You could get around this in a few ways depending on your dataset.

• You could adjust by the size of an area - say, looking at the number of incidents per square kilometre
• You could provide a rate per 1,000 occupants
• You could provide a % of the total occupants

Estimates of LSOA populations can be found online.

Using this information within SQL, Excel or (ideally) Python can help you to construct the most representative dataset and ensure you are not coming to incorrect conclusions.