2009/9/24 Jerry Clough - OSM <sk53_...@yahoo.co.uk>: > I've now had a few days to contemplate the Nottingham & Notts NaPTAN data, > so I thought I'd write > quick summary of what I've been doing. Feedback and suggestions all most > welcome. > > Use the data to name noname roads. I managed to resolve quite a number of > unnamed road issues thanks to the detail in the data. Using ITO's OSM Mapper > I've saved an image of the changes made based on this information: > http://wiki.openstreetmap.org/wiki/Image:Nottingham_NaPTAN_derived_road_names.jpg > Add major missing roads. Alignment of bus-stops have also allowed a much > smaller number of roads to be added. Being able to add these roads, which of > course will all need surveying, makes planning future surveys using only OSM > sourced data much easier. I use main roads for survey area boundaries, but > they are also very useful for estimating survey complexity (which even OSGB > 1:25000 aren't much use).
Nice work here. > I've merged very few bus stops as yet, just a few which I've happened to > pass. Overall impression is that the stops in the city of Nottingham are > accurately located. Some of the differences may be between boarding point > and the flag location (I always try and take this). All stops which I > encounter which are not already on OSM, where my GPS reading is close will > just be tagged as verified, and source=naptan_import;survey. > Details like indicator & shelter will not always be available immediately a > bus stop is verified. Today I passed a whole slew which I waypointed, but > I'd run out of convenient bits of paper and space on the dictaphone. I'd > rather revisit these than try and guess which had shelters/indicators. > Quite a few stops don't have indicators, but the infrastructure is being > installed. I propose to tag these electronic_indicator=installation. > One stop does not appear to exist on the ground: Charlbury Road ( > http://osm.org/go/eu8ZC@@J9-- ). A bus does seem to run along this road, but > I suspect it is a hail-and-ride sector. > > > Trams. I have been re-tagging tram stops to railway=tram_stop and changing > name to reflect signage at the stop (usually this means removing "Tram > Stop") from the name. As there is a stop for each direction, incorporating > the NaPTAN stuff for the NET-1 (Nottingham Tram) will take quite a lot of > work. Most of this needs to be done: the single node on the track did not > allow adequate representation of pedestrian access, particularly to island > platforms. I am proceeding to rework tram stops in Basford which are not on > a road. I am just creating two ways adjacent to the tram stop. One tram > stop, Cinderhill, has two stop points, but it has a single platform for up > and down direction trams, and a single track. This seems wrong, but I am > open to correction: particularly if the data model enforces such a > difference. This interests me, there should have been no tram stops included in this import. They were scheduled to be imported at a later date, with a lot of manual massaging from me beforehand. > Stop Points. Many stops belong to multiple stop points (presumably part of a > hierarchical arrangement), but some of these seem to have little utility. > For instance, stops BA82 and BA05 located at the junction of Vernon and > Nottingham Roads, belong to stop areas 339GBA09, 339GBA10 and 339GBA24. The > latter seems far too far away and is probably an error, but there seem to be > large numbers of overlapping stop areas, which to me as a passenger do not > seem logical. I'll look into this more once OPNV has rendered the data. > > Thanks to all who made this possible. > > Jerry Clough > > > _______________________________________________ > Talk-transit mailing list > Talk-transit@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-transit > > -- Regards, Thomas Wood (Edgemaster) _______________________________________________ Talk-transit mailing list Talk-transit@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-transit