Wednesday 12 November 2008 5:34PM
OK, it’s down to the wire. I’m presenting tomorrow. I was originally scheduled for next Thursday but I’m participating in an all-day workshop that day, so I switched to tomorrow.
I know what ‘d like my final form for this image database to be (I discussed before a 3-D tag cloud that shows the images), though I lack the skills to implement it overnight, but that hasn’t stopped me from searching for other people’s software projects that might come close. At first I thought of downloading the Thinkmap Visual Thesaurus to have a working demonstration of the navigational interface. Then I thought I’d look a little more to see if some open source project might fit the bill (and save myself US$39.95, though I’d likely use the thesaurus beyond this course – I check online dictionaries and thesauri semi-regularly as it is) and I found this Visual Complexity website that has collected an impressive array of software projects related to data visualization. A project called Flickr Graph (accessible via the Flickr Graph web interface) is getting closer to what I had in mind, except that it maps user relationships, not tags within sets or collections of photos… and for some reason it does not display my images in its grid (nor my Flickr user image).
Here’s a cool navigational interface for a repository of videos: TED Sphere (AKA VideoSphere).
Back on the Flickr tag track, I found TagGraph. The TagGraph web interface is similar to FlickrGraph but – oh yes! – it displays images and tags. Best part is that you can look at all photos for a given tag or, most relevant to my purposes today, by Flickr user – here is the TagGraph for ‘mikosian’ images tagged DM8305. If you click on ‘+ more images’ eventually all tagged images will be displayed. Then click on ‘+ related tags’ to see connections between images and various tags. This is where some of my tags (e.g. nyc) that are shared by photos outside the 100 Images set get pulled in, thus muddying the data pool (but not too much). I could go and remove tags on all other images to isolate the 100 Images set (since I can’t choose just that set through this interface), but that might be more work than necessary to illustrate a point (as opposed to actually implementing a digital photo archive navigational interface).
I had planned to create my own non-interactive (i.e. static) visual representation in Photoshop because the wireless network is so unreliable in our class room, but maybe now that I’ve found a working implementation of what I had in mind, screen shots will suffice. I only have to present for ten minutes, after all.
