search

Swirrl / ook

Structural search engine
https://search-prototype.gss-data.org.uk/
Eclipse Public License 1.0
6 stars 0 forks source link

Prune unused branches? #56

Open Robsteranium opened 3 years ago

Robsteranium commented 3 years ago

We might want to prune branches from codelists trees where none of the descendents are used.

This would mean users don't go browsing into branches where they'll never find anything usable.

It might also hide the encoding issues #55 in the NUTS data.

It may still be useful to show unused codes so we can say to the user "we know what you're talking about, the thing you're looking for does exist but there's no data on it". If the codelists were harmonised it'd be possible to conclude that "since I've search for Germany and the checkbox is disabled there must be no data about it". As it stands, they can only conclude "this version of Germany I've found has no data but I should continue searching to see if other definitions of Germany are in use". Thus, this loss of context is probably fine.

Robsteranium commented 3 years ago

@kiramclean I'd actually envisaged this as a UI feature (instead of having the ETL pipeline not import the data in the first place). It looks like you might already have implemented it! At least, I can't seem to see any of the mis-encoded geographies like Bulgaria in the NUTS codelists anymore (since resolving #45).