Earthify!

I’ve been spending a lot more time at Eyebeam recently, in large part due to the lack of work space in my current sublet. I’m on the lower east side now, in a room with no windows, where getting up at 11am feels like getting up at 4am. But I don’t mind my less-than-ideal living situation because I’ve come to enjoy spending every waking minute at Eyebeam. It’s a great environment with lots of creative geniuses and mad scientists, and since I work weird hours, I have the place to myself half of the time. I finally got some speakers set up with the Airport Express, so I can rock while I work (although I have yet to introduce this idea to the other fellows, so the rocking is limited to after-hours).

On Sunday, I was lucky enough to be at Eyebeam while a designer (who I later discovered was Banhaz Sarafpour) was introducing his stylish new Fall line of stewardess uniforms to an army of Style Magazine interns. I worked while listening to about 2 hours of test-runs featuring the same mix of Siouxsie’s Hong Kong Garden (the version in Marie Antoinette) and other fashion-friendly 80s remixes before I decided I needed a break to get some dinner. When I got back, the bouncer that Mr. Sarafpour installed at the front door apparently didn’t think I looked fashionable enough with my 5-day stubbeard and shaggy too-broke-for-a-haircut mop and tried to scare me away. But with a little perseverance, I was able to force my way in and finish up my new project:

Earthify

which is the only one of the three projects I am talking about here that is actually finished. Earthify is a script that takes any page of Craigslist posts – say, for instance, 1 bedroom apartments for rent by owner in NY that allow cats and dogs for under $1800 – and plots them on Google Earth. Devoted readers might remember that I came up with this idea about a month ago when I was actually looking for an apartment and was desperate for a way to browse available rentals on a map with a subway lines so that I could see what was convenient. Well, something amazing happened and I actually made this thing. There’s something to be said for safety-pinning that fruit on a low branch when you’re feeling unproductive.

It is currently running only on my local machine while I wait for Eyebeam to upgrade a server to PHP5, but I promise it will be available very soon and it will be glorious. In the future, I want to have Earthify scripts for lots of stuff, like shows, parties, the path that items on Amazon take before they make it to market. Stuff like that. It would be nice to have a generic Earthify script that spiders through a set of pages looking through addresses to map, but then how would you label the results? I think for now I will stick to site-specific implementations. I love the idea of a markup language for geographical data, and it amazes me that we already have such an impressive browser for such a new data format. Of course, Keyhole (the Grandaddy of Google Earth) has been around for a while, but its not often that a browser outpaces the development of the data that it was created to browse by quite so much. Whatever happens with Earthify, this won’t be my last project with Google Earth. KML is really a very simple but powerful markup, and I haven’t even gotten into the SketchUp stuff yet.

The Pretensotron

The Pretensotron (temporary name) is a text generator a la SciGen, only instead of computer science papers, The Pretensotron makes artist statements. This is a collaboration with Steve Lambert of the Anti-Advertising Agency, whose work you might have seen recently in the brilliant project “Light Criticism“. The Pretensotron asks the user for a name, a city, a medium, and a list of interests/concepts. Then it goes to work Frankensteining together an artist statement from the mutilated pieces of other statements and fits it all into a grammar developed by Steve and I.

Work on the Pretensotron has been underway for a while now. I have been developing a body of code for getting all of the information and analysis that I need. The Pretensotron is giving me the chance to explore more sophisticated text generation methods than the ones I used in Interactive Frank, and some new web services that I am excited about, such as Topicalizer and ClearForest. Of course, I am turning what should have been a quick joke project into a long-term, research-intensive one. But I’m getting a lot of stuff done here that I am going to use in my next project:

OPMLove

OPML isn’t one of those cool XML formats like RSS and FOAF that gets invited to all of the parties, so I’m just going to go ahead and tell you that it stands for Outline Processor Markup Language. If you use an RSS aggregator of any kind, chances are it will export an OPML file that includes all of the feeds to which you are subscribed. Also, chances are that if you’ve used an RSS aggregator for more than a few weeks, you are greeted with many more posts than you can actually, realistically read on an hourly basis. There are a few projects that have tried to remedy this problem, either with tag pools or image collages or summaries (including RS^3), but what I want to do is create something more layered and concentrated. My summarizer will create an audio segment that will summarize unread posts and convert the assembled text to speech, accompanied (perhaps) by whatever related media my army of hyperintelligent spiders can find. Kind of like what it must have been like for Keanu to have Kung Fu uploaded into that tiny brain of his in the Matrix. This will not necessarily be a pleasant experience – it will take every bit of concentration you have. But the goal is that, in the end, you will have a very clear idea about which stuff you want to follow up on and which you don’t care about. I’m even looking into some way that people could mark the parts of the audio that contains something that they are interested in, and then browse these items later.

People have a tremendous capacity for picking out phrases that interest them from a din of speech. I know this because I spend a lot of time eavesdropping on peoples conversations at crowded restaurants. And this filtering ability is significantly better than looking at a page of text and being able to pick out phrases that interest you. Tag pools are interesting, but representing an article with a single word isn’t always accurate. Collages are nice, but they require that all posts have images associated with them – images that also might not be accurate. With the audio collage that OPMLove will create, you will get a much more accurate idea of the content in your feeds in a much shorter time.

What particularly interests me about this idea is that it is confronts what would normally be treated as an information visualization problem, but attacks it rather as an information sonification problem.

Dirt Party

Finally, there’s the Dirt Party, the first official project of the The Savant Socialites Club for Thinkers and Adventurers, or SSCTA. The SSCTA is the semi-official name of the group that currently consists of Dave Jimison and I, which was formed as focal point for the work we have been doing, and a way to ensure that we keep working together despite our current geographical separation. But back to the Dirt Party.

The idea is that people must RSVP for the party, perhaps online, or maybe just sign up at the door. When they sign up, they will have to give their username for a few web services, like Flickr, LiveJournal, etc. The names are then fed into an Interactive Frank-like engine that will dig up photos, journal entries, audio, and video that relates directly to the people at the party and make a multimedia presentation that is generated and projected on the walls of the venue during the party. So at any given time, you might look up and see pictures of you or your friend flashing across the walls, or hear a computer voice reading a particularly personal journal entry that you wrote 5 years ago in some old forgotten blogger account. You will also suddenly become very close with the other people at the party – people you might not even know. It is the ultimate ice breaker.

We’re still thinking on this one and we will probably propose it at some festival this year.