We’re releasing Muddy in the next couple of weeks (as a web service complete with a range of APIs and also as an appliance you can rack).

It’s a service that indexes your content and finds ‘notable’ things and can tell you about those notable things (because the entities we identify are part of linked data).

This is cool for a heap of reasons:

• categorise your content dynamically according to the ‘stuff’ your content references
• create “aggregation pages” – aggregation of all your content around a concept (person, organisation, place, event) like, Madonna, or REM or Tony Blair or Shell etc. This is good for users and also SEO.
• produce “related links” – other indexing services do this but, well, we do it too.

All this supports the phenomena of Wilfing, browsing behaviour that stands for “What Was I Looking For”? Horizontal navigation (like related links) supports this ‘need’ amongst people in an age of digital content and this is how Muddy was born, but it offers the scope to do a lot more and one of those things is provide “informatics” about content. To explain this we pointed Muddy at an industry we’re increasingly doing quite a bit of work in, Museums. It’s also apt with Frankie, our Interaction Design Lead speaking and running a workshop at Museums and the Web (@mw2009) in Indianapolis this week (#mw2009).  Fortunately, the Powerhouse Museum last week announced that it was releasing its content under a CC licence (two different licenses are used but we were working on objects so were under the less restrictive license) so we thought it’d be nice to run their content through Muddy and see what popped out.

Of the 70, 000 (approx) items in the Powerhouse collection we sampled 10, 000 things. We then ran them through Muddy to see what was identified. We got quite a lot of positive matches, 50% or 5, 000 to be specific. We would have got far more but many of the objects in the collection have a very short description (this banana leaf cigar is one one we actually got – it’s an object from Samoa!) which does not give Muddy much opportunity to ‘find’ things. We thought location data was interesting and this certainly proved to be one of the most reliable entities that was returned. This is how the sample maps onto the world:

Click on the map for larger image. (We used Spatial Key to provide the mapping.)

This is kinda interesting because it starts to do two things:

• Enable you to analyse and tell stories about the collection
• Provide new ways into the objects in the collection – maps are not great navigational aids in our opinion but they are an aid.

We can see from the map that there is a strong Old World connection and particularly to the UK which will come as no huge surprise given the historical relationship between Australia (where the Powerhouse lives) and the UK. It could also suggest that the Museum is not as relevant to its region and also its minority populations as it could be.

Now this was just a sample of the entire collection. What would be interesting is to take individual categories from the Powerhouse collection and compare them to do a comparative study of the ‘types’ of objects. However, that’s not trivial given the nature of the categories presented. If we wanted to choose for example, electronics, electronics is a ‘silo’ that contains things with the heap of tags. One of those tags is games and games can contain non-electronic objects. So it’s not an exclusive taxonomy. Perhaps, if people think there are some specific tags that would be interesting to map we can do that.

So for us this is an example of Personal Informatics as applied to organisations. Organisational Informatics, producing representations of data that allow you to tell stories about an organisation in new ways, is something we’re really interested in as a way to give voice to objects and allow people to infer meaning and curate collections in new ways. We hope to be doing more of this kinda stuff in the coming months, once Muddy launches formally.

If you’ve got any suggestions on this piece of work (and organisational informatics generally) and what you’d like to see then please comment or message us.