Get hands-on experience with 20+ free Google Cloud products and $300 in free credit for new customers.

Feature Store Calculations

Hi,

I have a Google Colab notebook with some functions (Python) that been used to calculate the features for a model.

The functions use as inputs data from an API.

The question is if I can or should calculate the features inside a Features Store and feed the results to the Model?

Or in which Instance do I need to make the calculations and then feed the results into the model?

0 2 430
2 REPLIES 2

It is possible to use Vertex AI Features Store to Fetch the data, so you can use it as a part of the Vertex AI Workflow to train Custom or AutoML models in Vertex.

You can see here[1] the Vertex AI workflow.

[1]https://cloud.google.com/vertex-ai/docs/beginner/beginners-guide#workflow

Hi, thanks for your comments

After reading the documents, I understand that the data that should be
stored in the Feature Store is "static data".

By static I mean data from previously loaded databases and not calculated
within the Feature Store.

For example I wanted to add a simple average of the last 30 data entries
obtained from an API that sends data in real time in 5 minute intervals I
should::

Connect the API to the feature Store, store each data entry from the API
and then calculate the average inside the Feature Store?

Or should I connect the API to Goolge BiGQuery, store the data in Google
BigQuery, calculate the average and then send the data to the model deploy
in the end point?

Or connect a google colab notebook to the API, perform the calculations,
upload the Notebook to a container and send the data to the endpoint in
which the model was deployed?

Ance again thanks for your help