From raw data to interactive data app in an hour
summary
Claims an end-to-end data pipeline and interactive app from raw CSV to deployed model inside Snowflake using Snowpark ML modeling API and model registry. Workflow ingests a 54,000‑row diamonds CSV, does cleaning, ordinal and one‑hot encoding, min‑max scaling, trains an XGBoost regressor, runs grid search, improves MAPE from 0.07 to 0.06, logs the model, deploys for inference, and uses zero‑copy cloning and container services for GPU where needed. The demo runs and deploys successfully.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 5:19ML and pipeline challenges
- 9:11Snowflake vs data silos
- 13:27Unified Python workflows
- 17:46Snowpark ML optimizations
- 20:56ML modeling APIs
- 21:16MLOps model registry
- 24:26Setting up prerequisites and Snowflake trial
- 30:28Lab environment setup
- 35:11Data ingestion with Snowpark
- 46:17Feature engineering
- 58:26Training XGBoost
- 1:05:06Hyperparameter tuning
- 1:09:08Deploying models
- 1:14:03End-to-end summary
- 1:15:40Q&A
related talks
32mScalable Python and SQL Data Engineering without Migraines
29mBuilding End to End Gen AI Powered Analytics: From Data to Production
29mScaling Python-Powered Machine Learning With Snowflake
23mAcceleratinag Data Science and Machine Learning with Snowflake DataTech
01h 20mSupercharge Your Real-Time Analytics: Snowpipe Streaming & Dynamic Tables in Snowflake
34mPrepare data for machine learning with ease, speed, and accuracy
33mData Engineering simplified with Snowpark for Python on Snowflake
01h 13mBuilding a Semantic AI Layer in Snowflake: From Metrics to AI Insights