It looks a little ridiculous because of the buildings popping up all over the place. The hot new LIDAR data is pretty badass in some places like mountains, cliffs, and other sharp terrain features, but I seriously disagree with using it in urban centers. Don't get me wrong, it's super cool that they can show you altitude data of all the building tops, but think of it this way: how many people are walking around on the building tops? Maybe a few.
The vast majority of users are bound to the earth's surface: city streets and sidewalks. More importantly, if I'm walking or riding my bike across the city, the building tops don't affect my trip, only the street level elevation. So here's an idea: when you show me terrain on Google maps in urban areas, use the street level GPS elevation data from streetview cars.
OK, now take a look at these directions to San Francisco from Mountain View:
We all know the travelling salesman problem is one of the hardest around, but there are like a million better ways to get going on the highway in the right direction. It would be awesome for Google to provide an API so that aspiring data scientists could try out their best shot at this hard problem and compare results with the standard algorithm. How about that?
No comments:
Post a Comment