Machine Learning Using SQL?! — BigQuery BQML

Data scientists and data analysts can use BigQuery to analyze Big Data, fast, with SQL, no optimization, no Big Data code required. Wow. Did you know it can also run Machine Learning Algorithms?! Yes, AI with SQL, with little effort (our profile has free SQL tutorials). Yay for low-code AI. In a future post, we will discuss declarative AI, even less code. First of all, what is Google Cloud BigQuery? It is a great data analytics layer in the cloud that can connect to large datasets and query TB of data instantly. It is great for structured data, tabular data. It is not yet for unstructured data. In order words, it is great for data that fit nicely in tables. What also goes well with Tabular Data? Classical Machine Learning models. Any data analysts and data scientists who is familiar with basic SQL can create a machine learning model in BigQuery. Alternatively to machine learning from tabular data, you need working knowledge of Scikit-Learn ( or Tensorflow, Pytorch). Now you just need SQL.

Introduction to BQML — Datasets

BigQuery Google Cloud even provides free public datasets. Learn more here from this flash card. BigQuery dataset, table format, structure [public] include a link to the public datasets https://ml.learn-to-code.co/skillView.html?skill=vixd5O4D0hmmKJD7LbGY Developers can use BigQuery tables for building machine learning models. This means you can use it on your project or one of them many publicly available BigQuery datasets such as blockchain data and Reddit datasets hosted…