Clickhouse Integration

Integrated Machine Learning directly in ClickHouse

Make fast, relevant, predictions on the top of your data in ClickHouse using SQL. By integrating machine learning into your data layer, you can eliminate costly replication and simplify your application infrastructure.

Making Machine Learning native to ClickHouse

Your data lives in ClickHouse, so why do machine learning anywhere else?

With MindsDB we rethink the way machine learning models are trained and run. By moving  Machine Learning to the data layer you can.

Make fast, relevant predictions directly on top of your data in ClickHouse.

Eliminate costly replication

Simplify your application infrastructure

"The MindsDB team has come up with a creative way to integrate machine learning models with ClickHouse."

Let’s see how it works

The data

Used car sale price for Audi vehicles, the aim is to predict the list price when advertising a car for sale.
Source: Kaggle

Training a model

To train a model you just INSERT a new record inside the mindsdb.predictors table, which was created when connecting MindsDB to ClickHouse.​


To make a prediction you need to SELECT the model from the model table mindsdb.used_car_model and specify the appropriate features. i.e model is an Audi A6.

Now you can visualize your predictions in your BI tools

See how MindsDB works
with ClickHouse

Play Video