Last week Bill Jepson, Director of the Urban Simulation Laboratory and Chief Information and Technology Officer at UCLA School of the Arts and Architecture, showed off his Virtual LA at Harvard. (Luckily, a reader told me of the lecture or I’d never have known. Thanks Gary!)
The about 10, mostly Harvard Graduate School of Design (I think) folks, watched Jepson pilot us above and below the streets of LA examining full 3-D models of buildings, streets, subways (LA does have one) and the like. While we toured he told us how it all comes together:
With good CAD data and elevations it takes one person about 10 days to create a new building from scratch, that translates to about 60 hours per city block
newbies to the lab take about a full year to get up to full speed on adding to/changing the model
buildings and sections of the city are only updated when there’s a project (that is money paid) to do so
archetects’ models are not usually useable for addition to the model
the system is too big and complex to be licensed out, so all new additions go through Jepson’s lab
the whole model is terrabytes, but his unique binary compression makes it just gigabytes, that means among other things it loads in minutes, not hours
clients keep coming back, so he must be doing something right
He offered some interesting stories that highlight the big takeaway: this model is not for analysis (though there is a link to ArcInfo) but for human reaction. This one stuck with me:
At one point a professor at UCLA supported legislation that would require new tree plantings upon change of ownership of property. Jepson and his team used data from the past to predict what an area would look like “now” if such legislation was in place in the past. The trees were modelled such that they’d grow the right amount based on when they would have been planted. When the model was shown to the community, they were concerned about the decidious trees, suggesting that people would hide in them and jump down and mug people. The trees were changed to palms in the model and they were happy. (The legislation did not pass since realtors argued it would increase the cost of a real estate transaction.)
The other interesting story was about how this work in changing. What Jepson said sounded very much like what I’m used to hearing from Jack Dangermond about the evolution of GIS. It moved from large scale projects (small areas) to larger ones. It moved from modelling a single building to being a potential command and control center (with realtime data) for security and DHS purposes. It went from being theoretical to being practical. It went from being academic to having lives depend on it. Hopefully, it will also move from being so technical only specialists can use it to something more accessible.
A few things troubled me about not the model itself but about the social structure of it. Since so much of what geospatial technology is about these days is integrating data, I was disappointed this work while it did use raw CAD files, mostly seemed to recreate/create data. Jepson did explain that architects models are legal documents for construction; they are not meant for this sort of visualization. Further, he points out that every architecture firm has its own preferred “look and feel” and he needs to maintain a consistent one for his model.
Second, there seems to be no way for those of us outside this model and its team to participate, even by making models that might someday be useful in Virtual LA. For this sort of technology to really take off (pardon the pun), I think it needs to be less specialized. And, perhaps for now that’s not possible, but perhaps in the future?
Finally, this virtual version of a city made me think about what will happen as lay people craft 3D models in the free SketchUp and use them in Google Earth. Will it have any value? Any relevance to reality? Will we need trained specialists to craft these models so they are useful?
