Is It Possible to Run Axial Analysis on AutoDL for Large-Scale Road Networks?

[MenciNank]

I have a shapefile containing over one million road segments, and I’d like to run Axial analysis using this plugin and generate all the corresponding metrics. However, due to the limited computing power of my laptop, the software crashes every time I attempt to process the file.

I'm wondering if it’s possible to deploy this tool on AutoDL, a cloud platform with GPU resources, to handle the computation. The issue is that AutoDL doesn’t support QGIS GUI—it only supports running Python code.

Is there any way to run the Space Syntax Toolkit or at least the core axial analysis functionality in a headless (non-GUI) environment like AutoDL? Or do you have any suggestions for handling large-scale Axial analysis more efficiently?

[pklampros]

The core analysis functionality comes from depthmapX (depthmapXnet in the toolkit). If your online platform allows running binaries through Python then you can use the depthmapXcli (a command-line interface to depthmapX). Here are some examples: https://github.com/pklampros/depthmapWorkflowPython

Note: depthmapX (as depthmapXnet, or depthmapXcli or in any other form) does not use the GPU, so "a cloud platform with GPU resources" will not provide you with any speed up.

[jorgegil]

Hi @MenciNank , that is a great question.
SST was not intended for maps of that size because of the time it takes to process (with choice can take many days).
Furthermore, shape files are not the best for very large maps, since the large number of attributes added to the file can go over the shape file size limit of 2GB.

The processing engine behind SST is DepthampX
It has a command line interface, which should allow you to run your code on a server: https://github.com/SpaceGroupUCL/depthmapX/wiki/Command-line-interface

But it does not use GPU...

[MenciNank]

The core analysis functionality comes from depthmapX (depthmapXnet in the toolkit). If your online platform allows running binaries through Python then you can use the depthmapXcli (a command-line interface to depthmapX). Here are some examples: https://github.com/pklampros/depthmapWorkflowPython

Note: depthmapX (as depthmapXnet, or depthmapXcli or in any other form) does not use the GPU, so "a cloud platform with GPU resources" will not provide you with any speed up.

Thank you very much for your kind reply and suggestions!
I’ve carefully described my current issue and added all details (including my Python code and error messages) in this GitHub issue:
👉 #214

I would really appreciate it if you have time to take a look! Thanks again for your help and your great contributions to the community.

[MenciNank]

Hi @MenciNank , that is a great question. SST was not intended for maps of that size because of the time it takes to process (with choice can take many days). Furthermore, shape files are not the best for very large maps, since the large number of attributes added to the file can go over the shape file size limit of 2GB.

The processing engine behind SST is DepthampX It has a command line interface, which should allow you to run your code on a server: https://github.com/SpaceGroupUCL/depthmapX/wiki/Command-line-interface

But it does not use GPU...

Thank you very much for your kind reply and suggestions!
I’ve carefully described my current issue and added all details (including my Python code and error messages) in this GitHub issue:
👉 #214

I would really appreciate it if you have time to take a look! Thanks again for your help and your great contributions to the community.