i gave a keynote this morning at the Dublin Core Metadata Meeting - DC2008 on access to art museums on-line: a role for social tagging and folksonomy? that reports on more of the steve.museum tagging data analysis. this talk built on what i reported at NKOS last week [steve.museum: public and professional vocabularies. presentation @ NKOS 2008] and extended it to include some thoughts on user-generated metadata – useful in the context of DC, which began its life as a format for encoding user-created metadata – and a bit of work about the relationships between tags and search logs.

my slides are here (without some of the funky builds).

while we'd hypothesized that there might be a tight relationship between tags and search terms, what we found was a much looser coupling. whether this is a self-fulilling prophesy – because searches on the kinds of subject and genre terms that they use to tag fail, people don't use them – or because description and retrieval vocabularies vary at some other level still needs some thought. that's what the examples we looked at seemed to indicate, and a place i'll be looking further.

David and i presented the first of the steve.museum research results at the NKOS workshop today. the [many] slides are attached to this post. the take aways, though, can be easily summarized:

85%+ of tags are not found in museum documentation
60%+ tags don't match vocabularies [and those that do match ambiguously]
most tags can't be mined from other sources [like published catalogues or other scholarly works]

Public tagging vocabulary is different from the vocabulary in museum professional documentation. So tagging does contribute.

Contribute to what? well, we still need to look further into the details, particularly the relationships between tags and search terms to talk about that with more confidence. Watch for that from the Dublin Core (DC2008) meeting next week.

the fall travel season is starting up again, and we're in Europe this week and next presenting research results from the steve.museum project. watch for the following:

Public and professional vocabularies: comparing user tagging with museum documents and documentation

The 7th European Networked Knowledge Organization Systems (NKOS) Workshop at the 12th ECDL Conference, Aarhus, Denmark
Friday September 19th 2008 [see the program on-line]

David and i will be talking about the differences between public tagging vocabularies and the language of art cataloguing and curators.

and

steve.museum: tagging art. research and results

International Conference on Dublin Core and Metadata Applications, Berlin, Germany
Tuesday 23 September 2008 [see the program on-line]

i'm keynoting DC2008 – talking about the role of tagging in retrieval-focused museum metadata.

We're are looking forward to catching up with old friends, and – of course – i'll be posting notes and thoughts as we go.

The Indianapolis Museum of Art software development team has blogged some stats about an interactive installed in their Asian galleries, including some details about how and where it's touched. intriguing how, as Charlie Moad says, the heat map shows that people wanted to use the interface in the inverse way from how it was designed: they wanted to use the geographical map to find the work, not plot the work on the map.

our paradigms are shifting: "it’s google maps fault"

Some time ago – in 2005 – Archives & Museum Informatics launched a survey of museum web activities. We noted then that:

Each financial cycle museum Webmasters struggle
to justify their budget requests. Whenever statistical reports are
circulated someone asks, "How do we compare?" When exploring the
benefits of a new function, Web teams ask themselves "Is it worth the
investment?" Answers to these questions are hard to come by.

[see http://www.archimuse.com/research/mwbenchmarks/index.html for the full background]

This survey – the results of which are written up at http://www.archimuse.com/research/mwbenchmarks/report/mwbenchmarks2005.html  – was designed to help fill some of the voids in our knowledge. As a first stab at the problem we realise that it wasn't perfect. And things have changed since 2005!

We've been asked to update the survey for a client-group of museums, and thought we'd use the opportunity to ask for input from the community as well. If we're going to launch it again more broadly, we'd like it to be useful to you ...

So, please, take a look at the questions – available as a PDF from http://www.archimuse.com/research/mwbenchmarks/MW-Survey05.pdf – and post your thoughts on revisions or additional questions here. 

During the month of July, 2008, you can follow Netwurker Mez as she explores the artistic potential of Twitter. She's New Media Scotalnd's second Twitterer in residence.

Follow Netwurker Mez's tweets at http://www.twitter.com/mediascot

Anyone else doing fun twitter things?

/jt