Applying a created datamodel

[Kato]

I've created a random forest datamodel that predicts a certain parameter based on 10 different variables. Is there a way to apply this datamodel on live data and create a Seeq tag for the result?

Edited June 12, 2024 by Kato

[Patrick]

Hi Kato - Yes, you can apply your model on data in Seeq, although it won't be "live" in the sense of streaming results as new samples come in. 

What we usually do in a situation like this is to schedule the notebook using `spy.jobs.schedule()` to pull in the latest input data, apply the model, and then push the results back to Seeq. 

Check out the "spy.pull.ipynb", "spy.push.ipynb", and "spy.jobs.ipynb" notebooks in the "SPy Documentation" folder for examples.
 

[Patrick]

Hi Kato - I noticed you replied earlier with a follow-up question but it looks like it was removed.  I assume the above approach worked for you?  If not, let me know.

[Kato]

Yes, I managed to get everything to work using your tips! Thanks, Patrick 🙂

[Kato]

Hi Patrick,

I've realised my scheduling spy.jobs.schedule() often stops working. 

The error is: ScheduledNotebook 'NotebookJobs.NotebookJobs-0ef96d82-3bd8-7520-96b4-97106e5bd2a1_0' request was not run properly. HTTP status: 503, Body: {"canRetry":true,"statusMessage":"Could not start an instance of the Job because the container failed load or exceeded timeout of 60 seconds"}

How do I make sure my notebooks keep running?

Thanks!