Skip to content

Tutorial - LCLD

In this example, we will use the LCLD (LendingClub Loan Charge-offs) data to calculate the loan authorizing status.

Step 1: Logging in to the jupyterhub server

Access the jupyterhub server by navigating to the following URL: 

https://jupyterhub.{IDOML_DOMAIN}

Step 2: Cloning the example repository

Open a terminal in the JupyterLab interface. Run the following commands:

git clone https://github.com/serval-uni-lu/idoml-pipeline-example-lcld.git

cd idoml-pipeline-example-lcld

Open the Git graphical interface on the left side of the JupyterLab interface. Click on the clone repository button.

Image title

Paste the following URL: 

https://github.com/serval-uni-lu/idoml-pipeline-example-lcld.git

Step 3: Importing the dataset to the platform

In this example, we will use the LendingClub Loan Charge-offs dataset. The dataset can be downloaded by running the data retrieving notebook inside the cloned repository.

Step 4: Running the pipeline

Open the pipeline file lcld.pipeline in the JupyterLab interface. Then click on the run button to execute the pipeline.

Image title

Then select the IDOML runtime defined previously and click on the OK button.

Step 5: Checkout the pipeline execution

We can now check out the pipeline by navigating to the airflow server with the following URL: 

http://airflow.{IDOML_DOMAIN}

Note

If it is the first time you are submitting a pipeline to airflow, you will need to wait airflow to sync the DAGs. This can take a few minutes.

In the meantime, you can check the dag file in the repository which airflow is tracking for the pipeline execution.

Step 6: Check the Pipeline results

Once the pipeline execution is completed, we can check the pipeline results by navigating to the Minio server with the following URL: 

http://minio.{IDOML_DOMAIN}

Login with the following the user account, the execution logs can be found in the bucket defined previously in the IDOML runtime.

Step 7: Check the ML outcome

After the pipeline execution, we can check the ML outcome by navigating to the MLflow server with the following URL: 

http://idoml.mlflow.{IDOML_DOMAIN}