DEMO
Build a machine learning model using Snowpark on AzureML
In this demo, you’ll build a binary model based on the “Machine Predictive Maintenance Classification” dataset from Kaggle. You’ll also supplement this dataset with data from the Snowflake data marketplace. We’ll focus on information related to machine diagnostics, like torque and rotational speed, as well as environmental features like air temperature and humidity, to predict the likelihood of a machine failure. Machine learning models such as these are critical for verticals like manufacturing, where being able to predict machine failures can help inform timelines, cost, and operational efficiency. Here’s a brief look at some of the things you’ll do:
- Utilize Snowpark’s client-side APIs to load and manipulate data from Snowflake
- Train a machine learning model using Python in AzureML with ML Flow
- Deploy trained models in Snowflake using Python User Defined Functions (UDFs)