Scale of data

Background: I often need to produce thematic cartography for different applications (Environmental Impact Assessment is one of them). By law (Portugal), when you create thematic cartography for any official use, you need to specify the origin of the different data used, including scale/resolution information. I often need to mix different information, from different origins, with different scales (from OSM data to climatic data at 1:1.000.000...).

Vector data do have a 'base scale' or at least a 'positional error', that should always be considered when you are representing the data. It might be - or not - relevant to the use you are considering.

If the data was captured by a GPS unit, for instance, then you might have to consider a positional error corresponding to the positional error of the GPS reading. As an example, I have an old Garmin GPSmap 60C unit, which has a maximum accuracy of 4 meters, but which normally works with values of 6 to 9 meters, depending on the number of satellites available, their position on the sky, the degree of tree coverage, the presence of high buildings or structures...).

If the data comes from vectorized paper maps, then you need to consider the scale of the original map, which has its own positional error. In the Portuguese case, which is the one I'm familiar with, for 1:25.000 maps, the positional error is 12.5 meters, which corresponds to 0.5 mm on the paper. But if you have data digitized from a 1:1.000.000 map, then the positional error is much higher (500 meters, if we assume the same 0.5 mm on paper).

If you create cartography to be printed and used on the field, based on different data, it is fundamental to be conscious of their different origins and scale / positional error, and you should clearly state that information on the map regarding the base information. That is why I have been looking for the best way to describe OSM data scale / positional error.