[OSM-dev] Batch upload dilemma

Robert (Jamie) Munro rjmunro at arjam.net
Tue Mar 18 16:14:56 GMT 2008

Hash: SHA1

Andy Robinson (blackadder) wrote:
| Ciprian Talaba wrote:
|> Hi all,
|> One fleet management company was kindly enough to give us right of usage
|> to all their track logs, so now I am facing with the problem of using
|> this HUGE track database (the history from 2007 to now is something like
|> 40GB of data). One way I was thinking of using this information was to
|> have a daily archive of tracks (all the tracks, from all the cars, from
|> the previous day) and put them in an archive and upload it to OSM server.
|> The problem is that OSM server is treating these logs as a single track
|> and the timestamps from  different cars are mixed, so it is not usable.
|> The idea now is to upload each file generated daily by a car separated
|> using a script. I will create a different user for this so I can have my
|> own tracks separately. Do you have any other idea?
| Nice,
| Can I suggest you start a wiki page specifically to document your
| As Marcus has already suggested, some filtering of the data may be in
| too, though anything with an hdop of under 10 is probably useful in OSM.

I'd suggest anything with an hdop less than the distance to the nearest
road would be a good way to do it. That way, all points that are added
represent missing or inaccurate data. If there is a road that is missing
from the db and has only been driven down a few times by the fleet, it's
useful to know it's there, even if the hdop is large. If there is a
popular road, it's not worth keeping tracklogs of going down it unless
they are more accurate than the road in the database, so that someone
could use it to make fine adjustments to the position of the road.

Robert (Jamie) Munro
Version: GnuPG v1.4.6 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


More information about the dev mailing list