imports: de-duplication isn't always working

[zymurgic]

There's currently a number of imports that I've flagged as 'imported locations - awaiting approval' that had been 'imported and published' when it's fairly clear that there's already a location by that same charge point operator within 100 metres or so that has more complete information than the imported source, and in fact, reflects the exact same charge points - even if the Connections[].Reference may be different due to poor data quality in the imported data.

Hopefully, as various legislation improves, the quality of data held by organisations such as UK National Chargepoint Registry should far better match the exact current state of the devices operated by the charge point operators, rather than just what they keyed in when the charge point was originally installed - in some cases the Connections[].Reference has been changed due to business changes at the charge point operator.
https://www.legislation.gov.uk/ukdsi/2023/9780348249873

[webprofusion-chrisc]

Thanks Simon, I would gladly change all of these for OCPI data feeds direct from the networks if anyone can convince them to contribute data: https://openchargemap.org/site/about/datasharing

Imports like the NCR are kept around because they're all we've got, but if we can get something better quality then I'm all for it, as long as it follows some sort of open data standard - the days of us importing lots of different format sources are definitely over!

I did adjust the de-duplication radius recently for a different import because we were missing things on the other side of a freeway etc. Ideally the operators would clean-up the NCR data directly - it's their network their promoting after all and they have at least some financial incentive to get it right.