Evaluating correlation returning nan

[bappa10085]

I am trying to run the Molusce tool using rasters having no data

While running evaluating correlation returning nan

When I am running the Transition Potential Modelling, QGIS is getting hung.

The problem is in the backend MOLUSCE uses numpy Python library which cannot handle no data. So, how can I preprocess the spatial variables so that it can be handled by numpy Python library properly?

Please help me.

[RemeshevskiyValeriy]

Hello @bappa10085! Could you please send your dataset to my email remeshevskiy@gmail.com, it will help to better understand the cause of the problem.

[bappa10085]

@RemeshevskiyValeriy I have mailed you the dataset. Hope you have received that.

[RemeshevskiyValeriy]

Hello, Dear Dr Bappa Das!

Yes, I received your email. I'm sorry that I'm responding to you a little late. Thank you so much for reporting the error. Indeed, MOLUSCE incorrectly processes data that has the value NaN as "no data".

The spatial variable rasters from your dataset have a floating-point data type in which NaN is automatically used as "no data". At the moment, I can recommend you use rasters in which the value "no data" will be different from NaN. For example, you can use the Processing Toolbox -> GDAL -> Raster analysis -> Translate (convert format), in the tool you need to specify a new value "no data", and in the "Advanced parameters" specify the integer data type to which you want to convert the raster. Maybe you know more suitable raster transformation tricks in the context of your research, in that case, use them.