This is a tutorial for predicting credit risk (the risk of customers defaulting on loans). Your model will predict the probability of a given user defaulting on an outstanding loan.
The objective of this tutorial is to demonstrate how to work with the Online Hopworks Feature Store. The features are created with a batch program, but stored in the Online Feature Store. An online application then reads the precomputed features from the Online Feature Store to help predict if a given transaction is suspected of fraud or not.
This is a churn tutorial with the Hopsworks feature store, where you build a model to predict customer churn/retention. This tutorial also builds a user-interface with Streamlit that downloads the model from Hopsworks' model registry, and retrieves batches of data from the offline feature store for inference (making predictions). The model is trained with sklearn LogisticRegression.
This is an advanced example of the Hopsworks Feature Store usage; you are tasked with predicting the number of citibike users on each citibike station in the New York City. It is useful to compare Hopsworks' syntax with the equivalent tutorial for Snowflake/Snowpark.
This is an advanced example of the Hopsworks Feature Store usage. You are tasked with predicting the electricity prices in several Swedish cities based on weather conditions, previous prices, and Swedish holidays.
This is an advanced example of the Hopsworks Feature Store usage; you are tasked with predicting tomorrow's Bitcoin price using time-series data, and sentiment analysis for recent tweets about Bitcoin.
This is an introductory tutorial for the Hopsworks Feature Store. We introduce a fraud use case where a Python program loads data into two feature groups in the offline feature store. From there, you create a Feature View and training data with which you train a model to predict whether a given transaction is suspected of fraud or not. This is a batch use case, it will give you a high-level view of how to use our Python APIs and the UI to navigate the feature groups.
This is an advanced example of the Hopsworks Feature Store usage; you are tasked with predicting the Air Quality Index (AQI) in several European cities - Sundsvall, Stockholm, Malmo, Kyiv using weather features and air quality features for the previous days.
In this example, you write a PySpark program that produces and consumes messages to/from a Kafka cluster. This program can only be run from inside Hopsworks.
The notebooks in this linked folder (1) create features in Hopsworks and create a feature view, and (2) train a model and store experiment results in Weights & Biases. The model is trained using Scikit-learn, but is easily adapted to use TensorFlow, XGBoost, or PyTorch. The model output from the training pipeline is stored in Weights & Biases' model registry, but can easily be modified to store the model in Hopsworks' model registry.
The purpose of these notebooks is to give a brief introduction to Great Expectations concepts for data validation and for integration with Hopsworks. You can easily extend the set of expectations used on the sample data, and modify Hopsworks' policy for what to do when expectations fail - such as ignoring the failure or failing the pipeline.
This Python Hopsworks API example shows how to clone a git repository, checkout new branches to develop a feature, and push it to a remote.
This example of Python Hopsworks API program, shows how to create a job for an existing python or spark program, execute it and access application logs.
This example of Hopsworks Python API program shows how to create new projects or access existing ones in Hopsworks.
This Python Hopsworks example shows how to upload data to your cluster, such as for example a python program to run as a job. It also shows you how to download data from the cluster to your local environment.
This Python Hopsworks example shows how to run a Python program (from inside Hopsworks) that acts as an opensearch-py client for the OpenSearch cluster in Hopsworks.
This example shows the Hopsworks Python API for Kafka - manage schemas, topics and set up the connection to the kafka brokers.
This example Hopsworks program shows the Python API for managing secrets in Hopsworks.
This Python Hopsworks example shows how to interact with HopsFS from python, using the hdfs module.
This Python Hopsworks example shows how to run an example Airflow DAG that launches jobs on Hopsworks. This example assumes that you have already created four jobs in Hopsworks from the Jobs UI.