[geo-discuss] [Geodata] Re: Geodata in CKAN and collaboration (was Re: Responding to the consultation on opening Ordnance Survey's data)

Mon Feb 8 16:49:37 UTC 2010

Jo Walsh wrote:
> dear all,
>
> On 06/02/2010 13:32, Jonathan Gray [jonathan.gray at okfn.org] wrote:
>>    (i) If anyone is interested in joining a small informal group
>> focusing on CKAN's support for geodata and geospatially referenced
>> data - please drop me a line! 
>
> Also to note - I'm planning to draft a funding proposal on using open 
> geodata to link and integrate research data, which could take in 
> minimal metadata for CKAN packages of geodata, as well as OpenSearch 
> Geo, and involve several projects implementing the same interfaces.
Hi Jo - sorry I didn't get a chance to respond to this sooner. We've 
been busy opening up more data [1]

>
> Now I'm thinking of GeoCommons.com which has done a lot in the way of 
> minimal metadata and simple API search. How much free value can be 
> extracted from it or given back to it?
>
> I had a play around with the uploader. I'm wondering about the lack of 
> licensing clarity in GeoCommons.
>
> I uploaded a test shapefile that I found in a django test directory.
> There were options to add metadata describing where the data came 
> form, original source URLs, source authors and maintainers.
> I wasn't asked to add a description of data openness.
>
> In the boilerplate of the website there is a statement that "This data 
> is licensed under a Creative Commons Attribution 3.0 License".
> I wasn't asked to agree to this license while uploading data.
Thanks for pointing this out. What we actually mean by "data" in this 
case is the metadata. We don't mean to imply that you've modified the 
underlying data license.

We've been actively working with various groups and Lawyers on how CC 
applies to data. Obviously there has been a lot of work in ODbL, but 
we're still missing the "menu" of available licenses. Currently all 
licenses are unique unless they're using PD/CC0.

We could simply add a large text area for "copy & paste license here", 
but we're hoping for a more user-friendly option that then allows users 
to search across datasets, like one can do in Flickr, for "all 
redistributable" or whatever.

If there are suggestions for how to achieve this without making the 
metadata process daunting - definitely open to ideas. We want to help 
the global community make open data beneficial and easy to use. 
Consumption and demand drive more open-ness. See Nat Torkington's recent 
musings on the retrospective of New Zealand's open data initiative. [2]
>
> The upload app has a lovely second phase where you are prompted to 
> join this dataset to a number of in-built boundary files, or to 
> geocode it by picking the columns with addresses or placenames in, for 
> better instant gratification analysis visuals.
>
> However some of the sets of pre-cut boundary shapes are from gadm.org 
> - the GADM download page states "This dataset is freely available for 
> academic and other non-commercial use. Redistribution, or commercial 
> use, is not allowed without prior permission."
>
Good question - we've had the GADM data for quite awhile now and I 
believe permission was obtained. But I'll get clarity - and a good 
reason for some mechanism of specifying data openness in the metada  :)
> Now this is not the first time I have seen GADM redistributed under an 
> open license. But it would be good to be clear - there are some 
> applications in which data openness is not negotiable.
>
> Also, like Oscar, I would like to be able to search a geodata 
> collection at shape level, not just package level.

Andrew

[1] 
http://blog.fortiusone.com/2010/02/05/data-dissemination-to-the-government-of-haiti/
[2] http://radar.oreilly.com/2010/02/rethinking-open-data.html