I have discovered a problem with the data, which I would like to report:
The Garmin wanted to direct me in many locations in 90° turns down bridges from one parkway onto the other one. This is most likely because the highways in the TIGER data have common points exactly where they are crossing, although you cannot change the highway there (unless you want to some movie like jumps).
Is it possible to run a script over the TIGER data to check for those intersections?
Additionally many of the freeway lanes are not marked to be oneway only. Thus the garmin several times thought I was driving on the left lanes.
This is a common problem in most datasets. We used to call it a “highway jump”.
I don’t know if the original TIGER data has an attribute that defines the crossing roads as being on different levels. But if it doesn’t the only solution would be:
To join the OSM community and teach yourself to use JOSM
Split the node that connects the four road segments
Turn that position into two nodes that are a minimal distance apart. And re-connect the parallel road segments.
Have you considered participating in the OpenStreetMap Project to help correct these kind of faults?
I’ve talked about these problems in the Tiger data with Dave Hansen (who imported the Tiger data) because http://yournavigation.org also has troubles finding a good route in the USA. He confirmed that Tiger is missing this crucial info, because it’s purely topographic data: where the road is, now how it connects to other roads. So viaducts etc are not in the data. He is in the process of importing the Tiger 2008 data which is supposed to be much more accurate, but still does not include information about how ways are connected.
