How to integrate Machine Learning Model written in python with Anaplan modules
Hi Ana planners
I was curious to know how can we use/integrate a ML model in python notebook/script/pickle file within Anaplan modules?
Also if its not possible to integrate inside Anaplan modules then can some one suggest the structure of the process when the ML model is deployed in a cloud environment and the results are linked to Anaplan through APIs.
Answers
-
Love this question. Love it!
Not directly "within" a model but, yes, you can integrate with Python, no problem. You just have to use the Anaplan RESTful API set. I regularly pull data out of Anaplan, run it through my statistical modeling in Python, mostly for unsupervised ML since those generally require iteration on data. You can then import the results once your process is completed. My most common use case is clustering (k-means and/or hierarchical)
In addition, if you have access to a really good ETL tool, like Informatica, Mulesoft, or Dell Boomi, you can natively connect to Anaplan for the Import/Export portion. You'll need a good scheduler too.
Just remember, the data when loaded into Anaplan must be text or CSV format, so make sure Python deposits the results this way (do not use JSON).
Lastly, my personal favorite best practice article by @chase.hippen will certainly point you in the right direction.
Oh, you can also use R and Knime this way too.
2