The map really is not the territory

| Comments (2) | Misc
Apparently Google Earth has replaced the area photography of post-Katrina New Orleans with pre-Katrina images.
According to the GEC and my sources at Google, the imagery for New Orleans was actually changed last September. The previous imagery was directly after the storm struck, and was of inferior quality. Although the imagery of New Orleans is from pre-Katrina now, it is of better quality. If you have the Plus or Pro version of Google Earth you have the option to load two sets of post-Katrina imagery by logging out of the primary database. I think Google should consider getting more recent high quality imagery for New Orleans so it at least represents the present condition.

Apparently, Google selected a new set of high resolution photos for New Orleans. The only problem is that the new images are pre-Hurricane Katrina. So, all the damage that was caused by Katrina has now been erased in the Google Earth/Maps imagery database. CBS News says this move has sparked outrage and conspiracy theories in New Orleans. Ironically, the people in New Orleans have been some of the biggest fans of Google Earth as it helped save lives during and after the disaster. And, up until the recent update, residents used the pictures to illustrate damage to insurance adjusters, and to plan reconstruction efforts. Some of the conspiracies are that the local government itself requested the change to try and encourage tourism to come back to New Orleans.

Obviously, until the day that real-time satellite imagery is ubiquitous (probably not as far away as you'd think!) there's going to some tension between image quality and timeliness: is a timely but fuzzy image better or worse than a crisp but out-of-date image? While the answer does seem kind of obvious in this case and in other cases where the changes are dramatic and well-known, what about when the freeway on-ramp from my house is blocked this morning but the best images are from last week? It's not entirely clear to me that the modern fuzzy imagery is the right answer.

Current mapping and nav systems deal with this by treating maps as static and then overlaying meta-information (e.g., traffic, your directions), on top of the map. But if you had accurate remote imaging it might be more appropriate to simply display that—or maybe not. I certainly find it a lot easier to read traffic by seeing car density (and speed of motion) than the green and red lines on the Yahoo map displays, but there might be a display technique that would be easier yet. After all, maps are typically easier to get directions off than aerial imagery.


Mhm, well, there's "showing the wrong traffic" and then there's "showing a nonexistaent freeway". As you note, at least in this case, the answer should have been obvious... and someone made the wrong decision anyway.

I don't think anyone really believes that the images on Google Earth are timely enough to judge traffic on. There have been previous howlers, like their road map overlays crossing a river where the photo shows no bridge, or a series of roads spreading throughout what appears to be an open field, and that makes a better comparison: a high resolution image of a river area missing the connecting bridge, or a low resolution image with the bridge? Since the road information is available in the form of an overlay, maybe that's not so bad. In the case of a recently developed area? Well, maybe there a low resolution image is better than a high resolution image of dirt.

The dominating factor should always be utility. What benefit is there in a high resolution picture of dirt? A small missing element may not significantly detract from the utility of a map with many correct elements in place, but when the entire area is incorrectly displayed, one has to wonder why bother displaying it at all.

Well, they have to have some kind of policy on when to choose resolution over freshness. I am quite sure they will implement a function to allow choosing map sources sooner rather than later. It makes a lot of sense.

Personally I am not the least surprised. People would be shocked if they realized what kind of data a majority of their maps are based on. Many nautical charts are based on data from between 1800 and 1900 and on the fact that nobody has hit anything in the area since.

Leave a comment