I am trying to import meteorological forecast data into QGIS. Unfortunately the data lacks coordinate attributes, the spatial data is specified by the met institute as corner coordinates (SE, SW, NE, NW).

The data is in .JSON-format, but I managed to convert it into .csv as well. I tried to import as comma delimited text file but failed to add the georeferencing.

Any ideas?

Here's a link to the data if it helps:


  • 1
    The link suggests you have a point data set, not a polygon.
    – nmtoken
    Commented Nov 1, 2017 at 7:42
  • Have you a plan about how to handle that data in QGIS? The geometry is a single point and then there are a bunch of sensor observations from many time epochs. That does not suit to the data model of QGIS very well. Perhaps you could do something useful with QGIS by creating one point with the same coordinates for each epoch but that would require bit of scripting.
    – user30184
    Commented Nov 1, 2017 at 8:30

2 Answers 2


Unfortunately the data lacks coordinate attributes,

It doesn't.

The first line tells us:


The JSON file represents a set of timeseries data at a single point, as is hinted at in the URL to fetch the data:

geotype point lon(gitude)16 lat(itude) 58

  • Information on how to construct queries that return polygon and multipoint data are given on the following pages: opendata.smhi.se/apidocs/metfcst/index.html
    – nmtoken
    Commented Nov 1, 2017 at 8:32
  • I note that the API you are referencing is now deprecated: NB: As from may 2017 a new category of the API is released (pmp3g, version 2). This documentation covers that category and version. The previous category (pmp2g) will be phased out october 31 2017. Please upgrade to category pmp3g, version 2 before that.
    – nmtoken
    Commented Nov 1, 2017 at 8:34

Well this turned out far more complicated than i expected. As it seems, data can only be collected point by point. Super time demandning... :( Thanks for answering.

What I'd like to do is simply to visualize wind speeds and directions as a point layer in QGIS. The next step interpolate the data and use as input for an numerical wave forecast model. I guess QGIS could be a good tool for this?

  • The Swedish site seems to have good documentation and a working demo for gathering points from a larger area opendata.smhi.se/apidocs/metfcst/demo_multipoint.html. I fear that QGIS can do almost nothing out-of-the-box for you with this kind of multidimensional data. But weather data is not unusual at all and I am sure that you will find some ready-made and open source tools.
    – user30184
    Commented Nov 1, 2017 at 9:42
  • Could you please edit your question to include this new information instead of posting it as an answer? :)
    – Joseph
    Commented Nov 6, 2017 at 14:32

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.