From the course: AI-Powered Time Series Forecasting with Python

Unlock the full course today

Join today to access over 24,700 courses taught by industry experts.

Online prediction

Online prediction - Python Tutorial

From the course: AI-Powered Time Series Forecasting with Python

Start my 1-month free trial Buy for my team

Online prediction

“

Now that we know how to produce data to a real time stream and calculate our real time feature store, the last missing piece is to make predictions. So let's see how that works. First off, make sure your docker container is running from a fresh state. Go into the real-time folder and launch your docker container. Be sure that here on the left side there is no state folder. And then go to the data producer script and run it to produce some data. Once this is running, go to your feature pipeline and also start your feature pipeline so it's listening to the produced data. Now with that, let's close our terminal for now and go to the inference notebook. What we will do here is to define a Kafka consumer to subscribe to messages in the feature store. This one isn't using Quix Streams anymore, just Confluent Kafka. Let's take a closer look. The function fetch_all_feature_records is, as you might guess, fetching all feature records from our online feature store. The way this works is that we…

Contents