we are exporting ADM model data export, How can this data can be analysed externally ?
challenges facing:
Jason to csv conversion
pyModel data is encrypted.
Use performed export based on the below links guidance.
Exporting the ADM Datamart · pegasystems/pega-datascientist-tools Wiki · GitHub
Exporting adaptive model data for external analysis | Pega Academy
Exporting the ADM Datamart · pegasystems/pega-datascientist-tools Wiki · GitHub
Hi @SiddharthaB6233,
Can you please explain which method you have tried to export the data? Any method mentioned in the documentation on GitHub should work, if you face any specific issue with any of the methods we can try to help you out if you are able provide some more context.
With the encrypted column in point (2), do you mean the pyModelData column in the PR_DATA_DM_ADMMART_MDL_FACT table, as mentioned here? If yes, you can ignore the data in this column as you will not need it for any analysis.
Best,
Tom
@SiddharthaB6233 To analyze exported ADM data externally, first convert the JSON data to CSV using tools like Python or jq. In Python, you can write a script to load the JSON, extract keys as headers, and save the records in CSV format. If the pyModel data is encrypted, you will need Pega’s decryption API or utilities. Work with your Pega admin team or support to access these tools or enable unencrypted export by modifying data extraction rules, ensuring compliance with security policies. Once converted and decrypted, clean the data for accuracy and use analysis tools like Excel, Tableau, or Python libraries to visualize trends. Always store sensitive data securely and seek Pega support if issues persist with decryption or export processes
@SiddharthaB6233 If you have additional questions after Tom’s answer please let me know. You can also post to the PDS tools discussions: pegasystems/pega-datascientist-tools · Discussions · GitHub.
You should not attempt to create CSV, instead just export the dataset to the standard Pega format (which is a flavor of zipped NDJSON). As Tom mentioned, the pyModelData field is internal and you should be ignoring this in the analysis. It could (if size is a concern) be dropped from the export as per the article mentioned by Tom.
-Otto