Incremental Spatial Autocorrelation: Restore the PDF output (Global Moran's I) feature

[wayfaringrob]

In recent versions (currently observing on 3.3.0), Esri has removed functionality of the Incremental Spatial Autocorrelation tool. It used to spit out a PDF containing a Global Moran's I table and chart. Now, this is only available in a database table or the messages, which I find quite clunky and less useful than the standalone report. Copying/pasting values from the table in messages is especially maddening (what do you mean "copy all?")

In addition, Esri's own documentation still mentions the PDF output: https://pro.arcgis.com/en/pro-app/3.3/tool-reference/spatial-statistics/incremental-spatial-autocorr...

But this parameter is missing from the GP tool.

Further--and I understand this is a niche use case--I would like to have students submit this file with their homework. The "improved" method makes this all harder.

Please restore the PDF output from this tool. Thank you!

[DanPatterson]

BUG-000170703 for ArcGIS Pro

as a followup and potential workaround for some

[AnkitaBakshi]

Hi @wayfaringrob ,

The parameter was intentionally removed from the UI due to the limitations associated with the PDF report generated by the python library matplotlib. This library has compatibility issues with certain languages and newer windows operating systems, which can result in missing fonts and consequently an empty PDF. In summary, relying on this method is not a sustainable solution to ensure the PDF report works effectively for all users. Instead, the new charts and geoprocessing messages are more standard and recommended ways to access the information. However, you can still create reports using arcpy. If you run the tool using arcpy, you can still access the PDF output parameter. 

Since the parameter is there in arcpy, the documentation still has it. It is only hidden from the UI to encourage the use of charts and geoprocessing messages. We have also now added a note about it in the toolbox history for the next release.

[wayfaringrob]

@AnkitaBakshi  thanks for the intel. If the functionality is still there, why make it so much harder for users who need it and don't have any issues with it?

[AnkitaBakshi]

@wayfaringrob PDF is not the recommended way to access the information anymore because of the issues mentioned above. Therefore hiding the UI parameter encourages the use of geoprocessing messages and charts which is recommended moving forward. We kept the parameter accessible in python to not break existing workflows. Hope that helps clear the confusion. Thank you!

[wayfaringrob]

@AnkitaBakshi  hold up. If the rationale is to keep things accessible so as not to break existing workflows, then why has the GUI parameter gone missing? How is that not breaking an existing workflow? GUI to PDF is a workflow!!! This is a needless and infuriating removal.

[wayfaringrob]

@AnkitaBakshi  you also previously stated that the "only" reason it was removed was to encourage using messages--which seems to imply that there is no real problem with PDF. Has anyone complained about having easy access to the PDFs? They're optional after all, I'm just extremely perplexed by what the actual issue is here. Can you shed any light on the problem with PDF?

It is irritating that Esri will let bugs fester for years and tell people to deal with it, but then removes things that are used and loved by its customers. I really dislike using messages in Pro. They are a hassle to view, open, and save. PDF was perfect. I wish EVERY tool could make a PDF or some kind of external log of itself. Way less hassle. Way more flexible. Might there be formats other than PDF it can produce without the errors you allude to? I would accept separate image and table files, even (default for table should be changed to outside .gdb for tools like this anyway--distance bands don't really have a reason to be there). But messages are a less than ideal solution unless the UI is drastically improved and export functionality is built in. The solution shouldn't be to give up on PDF because it's giving issues now; it's to fix the export format so it works. Please reopen this idea. We're giving up too easy.

[wayfaringrob]

@AnkitaBakshi  also, this is the recommended workflow in Esri's own articles on the topic: https://pro.arcgis.com/en/pro-app/latest/tool-reference/spatial-statistics/choosingdistanceband.htm

Spatial Autocorrelation spits out an HTML document -- would that be a possibility here too?