[okfn-labs] Crowdcrafting + reference lists = crowdsource lists of stuff?

[Jonathan Gray]

Thanks Tom. I don't think that organising users would be an issue, as we'd
likely be looking at fairly small tasks - usually related to projects 'with
a cause' (like investigating climate denial groups). Part of the point
would be to find a way to engage people around these projects, not to use
existing mechanical turk projects to pay people.

My point was less about the specific case of finding Twitter handles but
more whether PyBossa could be generalised to deal with something that I
feel might be a recurring problem.

Just a thought!


On 21 February 2013 18:47, Tom Morris <tfmorris at gmail.com> wrote:

> On Thu, Feb 21, 2013 at 10:26 AM, Jonathan Gray <jonathan.gray at okfn.org>wrote:
>
>> Agreed that canonical reference data should be authoritative rather than
>> opinion based. ;-)
>>
>> I'm talking more about using standard reference lists (e.g. counties in
>> the world, cities or counties in the UK) to then generate new lists which
>> may not be 'reference data' per se - like 'Twitter account for every local
>> education authority' or 'URL of national contact point for multinational
>> corporation X'.
>>
>> Generally this would be a way to get a piece (or several pieces) of data
>> for every item you had on a given list. A recent use case would be wanting
>> to get the Twitter accounts for a given list of climate change denial think
>> tanks [1].
>>
>> What do you think? Useful/not useful? Perhaps Google Docs is the easiest
>> and best way to do this for now?
>>
>
> Microtask sites are used for this kind of work all the time and you can
> get it done pretty cheaply.  In additional to Twitter accounts, people use
> this technique for lists of home pages, customer service phone numbers,
> etc, etc.  The disadvantage to using Crowdcrafting specifically is that it
> has no user base, so you'd need to organize your own volunteers, whereas if
> you used Mechanical Turk, Crowdflower, etc you'd have a waiting army of
> people to do your work.  You would need to pay them, but only pennies.
>
> For your specific task of finding Twitter handles for organizations, I'd
> first see if I could get a list of home page URLs using Freebase, DBpedia,
> or Wikipedia and then set a scraper running on that list of URLs to see how
> many Twitter handles I could scrape automatically.
>
> Tom
>



-- 
Jonathan Gray <http://jonathangray.org/> | @jwyg <http://twitter.com/jwyg>
Director of Policy and Ideas
The Open Knowledge Foundation <http://okfn.org/> |
@okfn<http://twitter.com/okfn>
Support our work: okfn.org/support
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/okfn-labs/attachments/20130221/973e173f/attachment-0002.html>