Skip to content

Test a "floor number" feature based on unit numbers #90

Description

@jeancochrane

Roughly 50% of condo PINs have unit identifiers that start with a number:

select numeric_unitno, count(*) as count
from (
    select
        case
            when
                unitno is not null
                and regexp_like(unitno, '^[0-9].*')
            then 'yes'
            else 'no'
        end as numeric_unitno
    from iasworld.pardat
    where cur = 'Y'
        and deactivat is null
        and class = '299'
) as unit_counts
group by numeric_unitno
order by numeric_unitno desc
#	numeric_unitno	count
1	yes	        6937526
2	no	        3501778

It's likely that for many buildings, particularly high rises, these numeric unit identifiers may be structured and contain information about the floor that the unit is on. We should try making a "floor number" feature based on unit identifiers in cases where the identifier appears to be numeric and follows the common pattern of a floor prefix with a unit suffix.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    new data/featureCreate or edit a column/feature or collect new data

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions