[ckan-dev] Spatial Extension : Datasets without spatial metadata

David Read david.read at hackneyworkshop.com
Wed Jul 18 10:20:23 UTC 2012


Hi Oceane,

Great to hear the spatial matching would be helpful too. We might get round
to it if OKF don't beat us to it.

The change in the postgis query sounds like a special customisation, so
you'll probably need to ask a python developer to work on it. You could
always try contracting the OKF or one of the CKAN partners if you don't one
available at ATOS.

David

On Wednesday, 18 July 2012, VENTURA Oceane wrote:

>  Hi, ****
>
>
>  I understand the reasoning. Indeed, calculation of the similarity of two
> spatial areas can be a very good feature.****
>
>
>  Meanwhile, I prefer to avoid having too large extents and keeping the
> spatially-undefined datasets (for a better definition later). How can I
> modify the Postgis query?
>
>
>  Have a nice day, ****
>
>
>  TIA, ****
>
>
> *Océane VENTURA*
> Business Development & Innovation
> ATOS - Public Sector Franc
>  *oceane.ventura at atos.net <javascript:_e({}, 'cvml',
> 'oceane.ventura at atos.net');>*
> River Ouest,
> 80 quai Voltaire
> 95 877 Bezons Cedex
> www.atos.net
>
>
>   ------------------------------
> *De :* ckan-dev-bounces at lists.okfn.org <javascript:_e({}, 'cvml',
> 'ckan-dev-bounces at lists.okfn.org');> [ckan-dev-bounces at lists.okfn.org<javascript:_e({}, 'cvml', 'ckan-dev-bounces at lists.okfn.org');>]
> de la part de David Read [david.read at hackneyworkshop.com<javascript:_e({}, 'cvml', 'david.read at hackneyworkshop.com');>
> ]
> *Date d'envoi :* mardi 17 juillet 2012 15:27
> *À :* CKAN Development Discussions
> *Objet :* Re: [ckan-dev] Spatial Extension : Datasets without spatial
> metadata
>
>  Oceane,
>
>  If a dataset doesn't have an extent defined then I would agree with the
> default of excluding it from all spatial searches. If nothing else, it
> encourages publishers to define a datasets' extent.
>
>  If you really wanted to include spatially-undefined datasets, then you'd
> have to change the Postgis query.
>
>  In data.gov.uk we also have the opposite problem: many datasets have the
> extent set to the whole of the UK (because of laziness/default), even if
> the dataset applies to one city. So you search for London and datasets from
> Edinburgh are in the results, which is perhaps just as annoying as not
> getting any results.
>
>  An improvement would be to weight higher in the search results those
> datasets with a spatial area 'similar' to the search area. So a search for
> London datasets would rank London datasets higher than UK datasets. I
> wonder if @Adria knows a recognised way to calculate the 'similarity' of
> two spatial areas to achieve this?
>
> David
>
> On 17 July 2012 10:21, VENTURA Oceane <OCEANE.VENTURA at atos.net> wrote:
>
>  Hi,
>
>  I just installed the spatial extension and I noticed an unexpected
> behavior when I do a research.
>
>  With the spatial query widget activated, if the spatial extent is not
> defined, there is no result. If I specified a geographic area, CKAN find
> the package with spatial metadata.
> But I never have the package without these metadata in the result.
>
>  Is it possible to do the research on package without spatial if the
> extent isn’t specified by the user?
>
>  TIA,
>
>
> *Océane VENTURA*
> Business Development & Innovation
> ATOS - Public Sector France
> 01.73.26.47.92
>  06.32.33.25.99
>  *oceane.ventura at atos.net*
> River Ouest,
> 80 quai Voltaire
> 95 877 Bezons Cedex
>  <http://www.atos.net/>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/ckan-dev/attachments/20120718/40fe1170/attachment-0001.html>


More information about the ckan-dev mailing list