I'm not particularly well versed in GIS, so maybe the answer is obvious:

I have data provided in textfiles on an FTP server here. Each column corresponds to a value on a grid laid out over germany. Each cell in the grid has a size of 1 square kilometer and the projection used seems to be EPSG:31467.

If I check the projection here, it doesn't seem to cover the whole area covered by the data set provided. Why is that? How would I generate the correct vector grid (fishnet?) in QGIS to correlate it with the data?

  • 1
    You don't need to create a grid, just drag and drop the unzipped .asc file into QGIS, and set the CRS of the layer to EPSG:31467. It does not matter that DHDN zone 3 is actually defined for a part of Germany only.
    – AndreJ
    Commented Jul 25, 2017 at 14:46
  • One question per question please. Right now you have one about the projection and one about making a grid in QGIS. Commented Jul 25, 2017 at 14:46
  • The spanning question would be "How does one correlate german cdc-data with a 1km x 1km grid?" I guess, the other questions just being pointers into what I would have guessed to be correct answers.
    – heyarne
    Commented Jul 25, 2017 at 14:57
  • Correlate how? Do you have that other grid? Do you want to do spatial analysis or just transfer attributes? Commented Jul 25, 2017 at 15:00

1 Answer 1


This projection does not cover the whole region as it normally would lead to large distortions. This projection keeps angles untouched but therefore distances differ up to 1m per 1km (0,1%) in the normal usecase of the projection. This differences depend on the distance from the central meridian defined in the projection. Therefore you define several "zones" to avoid larger distortions. For Germany wide usage you often only have one projection to avoid having several datasets (that would not touch each other and having a gap to the neighboring dataset).

The distance deviations should still be in the range of a few meter per kilometer and should be no problem for your dataset. The data is interpolated to the center or top left border of your grid cell and will therefore have much larger deviations in the data itself. But of course you have to keep this in mind when correlation to other data.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.