Thursday 4 February 2010

Crowd forces

Crowdsourcing is a short-hand term used to describe putting out a task that would have usually been performed in-house to a large group of people. Sites like Threadless, and tools like the Brooklyn Museum's 'Tag!' game epitomise how crowdsourcing can work beautifully on the web.

Today I noticed two new crowdsourcing projects.

First I saw a tweet from Christchurch Art Gallery, talking about the way their Friends group is geo-tagging their collection. As far as I know this is not being done online, but the ethos holds - a group of people donate their time & knowledge to perform a task that would usually have been done by a staff member (but which might never have been a high enough priority to get on to their work plan).

Of course, volunteer projects inside galleries and museums are nothing new. But it's unusual to hear about them, and for me at least, unusual to see the benefit. I'm guessing the combination of a regionally-focused collection and a supporters group who lives in the region is going to help make this tagging very accurate. You can see the tagged collection items here - each full record has a link to a Google Map at the bottom.

Also today - the V&A launched a beta crowdsourcing site, this one designed to get the public to help identify the most pleasing crops of photos of collection items to use in their online browse.

Obviously this is a lot more complicated than what Christchurch is doing. It's also a little buggy around the edges (my tally fell off part way through my session, so I couldn't see how many images I've completed) and take a few goes to get the hang of - you have to make several choices related to each item before it's 'completed', but the signals that tell you this are hard to pick up.

Unlike the Christchurch example, I wonder how accurate this will be. When I started making choices, I realised I didn't really know what criteria I was meant to be using. Did they want the colour scale in, or out? Did they want as much of the item as possible, or a great detail? In the case of the item below, what *is* better - side on or bird's-eye? What if I thought a crop looked great, but it was better photo than it was a representation of the item? And did anyone agree with the choices I was making?

Hopefully the V&A is following the release early, release often mantra, and will keep tweaking the in response to the feedback they're getting. It's a great concept, and deserves that.


Tim Jones said...

Thanks for your kind words. In fact accuracy, or rather what we even mean by accuracy, is important in the Christchurch project too. Are we tagging the subject? or the artist's viewpoint? The middle of Lyttelton Harbour? Or the shoreline that is in the foreground? What about elevation? What if the building depicted no longer exists? Or if the view can no longer be seen?

Using a blog to debate the contentious ones might well be our next step.

And we are not just tagging local scenes: we have identified many sites and buildings in the UK, France, Italy, Antarctica and even India, but we would obviously welcome further local knowledge.

Tim Jones - Christchurch Art Gallery Te Puna o Waiwhetu

Courtney Johnston said...

Thanks Tim!

I hoped that I indicated that I thought you were doing a terrific job on the accuracy front. I do the geo-tagging for the photos the Library adds to The Commons on Flickr, and often find myself facing these questions especially the subject/photographer's position. Or, have you ever tried to find a particular point in Antarctica using Google Maps?

I wonder if the next step after accuracy is entertaining non-accuracy - like, can you do something fun with aggregate landscapes, like Angus's Central Otago, or imaginary places?

paul said...

And now they are using the mapping this to act as a companion to the new Sutton show - showing the places he painted on his tour of Italy on a map.

Simple - but cool.