Recent searches
Search options
Administered by:
If you're trying to teach people about the horrors of data processing and attempting to normalise user input, may I suggest you get them to look at the "country_of_residence" field in the companies house Persons of Significant Control data file?
People can read and write well enough to setup a company but not well enough to be able to spell "United Kingdom"
I think my favourate one so far is "United Kingdum"
Special shout out to the 438 people who filled out the "Country of residence" field with "London"
And the 64 people who put "United Kindgom"
ok, a new winner "United Kinkdom" entered by five people. A small union of dominatrixes?
@emily_s what's the emoji for jaw hanging slack?
@ryanc I'm not sure, if there is one it doesn't have sufficient gravity for this mess
@emily_s some proud cockneys there! i'd wondered if companies house data came from well validated gov.uk style forms, but this data quality suggests otherwise - ps: fun stream of real-time companies house data https://companies.stream
@joeldn I'd be tempted to guess that a lot of this could have been OCR'd from handwriting given how bad it is. Either that or there is no validation on this at all.
Even with the cleaning I've done so far I've still got 2385 distinct values for country of residence. Though I've not started sorting out the town names yet. All the oxfords and readings
