Hi all. I am a newbie developer and am just getting my feet wet by parsing the OSM database. 130GB of XML text was a fun challenge.
I have been using the Tiger database. The 2008 Tiger data looks really good except where it is totally messed up due to some massive database error over at the US Census. Where it is not messed up I notice that it can be a lot more detailed than the OSM data in the same area. The road I live on has at least 30 points in Tiger and 2 in OSM for example. As I understand it, OSM in the US was originally based on the Tiger data. Does this mean that the Tiger database has since be greatly upgraded or did the process of importing Tiger into OSM cause a lot of data to be lost? Is there any plans to use the higher precision Tiger data in the future (like after they have fixed the data base problems in 2009 and 2010?)
More points doesn’t make things better, it can be redundant data, can you give an example of what your describing?
I was in the process of locating an online Tiger map data viewer when I answered my own question. The Tiger viewer I located on the tiger.census.gov page displays 1998 Tiger/Line data. It precisely matches the OSM data for the area around my home. This means that OSM is based on an earlier version of the Tiger data.
As far as your point that more points can just be redundant, I do agree. When I was processing the 2008 Tiger data for my own project, I had to cull a lot of points that were not really needed for the map to look fairly smooth. I suspect that the cartographers at the US Census (or wherever they are buying their data) use spline curves to model the roads and then convert these into polylines for export to shapefiles with a fairly high precision tolerance. However, in addition to simply having more points, the 2008 Tiger data is also a lot more accurate. There are notes on the Tiger site that explain that over the past few years they have been working to greatly increase their maps accuracy and it really shows. Except for the annoying data corruption I mentioned in my first message, their maps look spectacular.
I think this may also explain why OSM is missing most of the streams in the 2008 Tiger data. I have been very pleased with the large number of streams and small water bodies that are in the Tiger data set. There are very few in OSM.
I think it may be worth considering creating a way to re-merge the current Tiger data into the OSM data since it has much better precision. Even if you only added the streams and water bodies from Tiger, it would greatly increase the quality of the US OSM data. There may also be a way to compare roads and replace matching roads in OSM with ones from Tiger that don’t diverge too far. This would take care of most of the Tiger data corruption issues. I am thinking of merging OSM and Tiger for my own mapping project by taking the roads from OSM and the water from Tiger.
This forum doesn’t get a look in much by the powers that be, you are best off joining the main OSM talk list to discuss such things.
Will do! Thanks for responding.