Set up a Looker project

MDE provides a Looker Project Template git repository to help you bootstrap a new MDE Looker project. We also provide a library that includes MDE-specific LookML base components and starter configuration packages to reduce the time it takes to model your MDE data in LookML.

Before you begin

To complete this guide, you need the following prerequisites:

1. An active Looker instance.
2. A Looker user account with the Admin role.
3. An active MDE deployment in a Google Cloud project with the default MDE configuration package applied.
4. A user account with at least the following Identity and Access Management (IAM) roles for the Google Cloud project referenced:
   • roles/iam.serviceAccountCreator
   • roles/iam.serviceAccountKeyAdmin
   • roles/resourcemanager.projectIamAdmin
   • roles/source.admin
5. Workstation with a local installation of:
   • git
   • Google Cloud CLI

How to bootstrap a new LookML project

1. Set the default gcloud project to your MDE deployment project.

   All gcloud commands in this guide assume that the default project is set to the MDE deployment project. You can set the default project using the following command:

   gcloud config set project PROJECT_ID
   

   Replace the following:

   PROJECT_ID with the MDE deployment project ID.

2. Create a service account that Looker uses to read data from BigQuery, and generate a JSON key for the created service account using the following commands:

   export SA_LOOKER="mde-looker"
   export PROJECT_ID=$(gcloud config get-value project)
   gcloud iam service-accounts create $SA_LOOKER --display-name "MDE Looker account"
   gcloud projects add-iam-policy-binding ${PROJECT_ID} --member serviceAccount:${SA_LOOKER}@${PROJECT_ID}.iam.gserviceaccount.com  --role "roles/bigquery.dataEditor"
   gcloud projects add-iam-policy-binding ${PROJECT_ID} --member serviceAccount:${SA_LOOKER}@${PROJECT_ID}.iam.gserviceaccount.com  --role "roles/bigquery.jobUser"
   
   export SA_LOOKER_KEY=~//mde-projects/$PROJECT_ID/${SA_LOOKER}_key.json
   gcloud iam service-accounts keys create $SA_LOOKER_KEY --iam-account ${SA_LOOKER}@${PROJECT_ID}.iam.gserviceaccount.com
   
   cat <<EOF
   ===========================================
   Connection credentials to be used in Looker
   ===========================================
   
   ConnectionName: mde-bigquery-connection
   ProjectName: ${PROJECT_ID}
   Dataset: mde_data
   ServiceAccount Email: ${SA_LOOKER}@${PROJECT_ID}.iam.gserviceaccount.com
   ServiceAccount JSON key location: ${SA_LOOKER_KEY}
   
   EOF
   

3. Create a new BigQuery connection in Looker with the following specifications:

   • Name the connection as specified in the output of Step 2.
   • Set the dataset as specified in the output of Step 2.
   • Follow the service account authentication path, and use the service account key available in the location specified in the output of Step 2.
   • Configuring PDTs for the purposes of this guide is unnecessary. You can always configure PDTs for this connection on a later stage.
   • Set the database timezone to UTC.

4. Create a new git repository. This repository will host the LookML code of your Looker project.

   The following command creates a GitHub repository using the GitHub CLI. However, you can create a repository with any supported git service.

   gh repo create mde-looker --private --clone
   

5. Navigate to the newly created repository it using the following command:

   cd mde-looker
   

6. Pull the MDE Looker Project Template files, and push them into your new Looker repository with the following commands:

   git checkout -b master
   git remote add project https://github.com/GoogleCloudPlatform/mde-looker-project-template.git
   git pull project master
   git push -u origin master
   

7. Create a blank Looker project.

8. Connect the project you created in Step 6 with the git repository you created in Step 4, and register the SSH key that Looker generates with Cloud Source.

9. Configure access to the MDE LookML Library.

   The MDE Project Template declares the MDE LookML Library as a dependency. Since the MDE LookML Library is hosted in a private GitHub repository, the public SSH key of your Looker project must be allow before you can access it:

   1. Obtain the SSH key of your Looker project.

      Configure authentication for the MDE LookML Library using the SSH method and obtain the public SSH key of your Looker project.

   2. Request access to the MDE LookML Library. Once you have obtained an SSH key, fill out the MDE LookML Library Access Form for allowlisting. Your Looker project will have access to the library after the SSH key has been allow.

   3. Test access.

      After the SSH key of your Looker project has been allow, you can test that your Looker project can connect to the GitHub repository where the MDE LookML Library is hosted.

10. Open the manifest.lkml file in Looker, and supply the required values:

    • Replace <TAG_METADATA_BUCKET_NUMBER> with the tag metadata bucket number.
    • Replace <SOURCE_METADATA_BUCKET_NUMBER> with the tag metadata bucket number.

    You can look up the bucket numbers in the MDE web interface or by calling the GET/metadata/v1/buckets/tag and GET /metadata/v1/buckets/source API endpoints. Alternatively, you can look up the bucket numbers in the MDE UI or Manufacturing Connect.

11. Congratulations, you have set up Looker for MDE.

    The LookML model in MDE Looker Project Template imports the LookML explore file and dashboards from the default configuration package in the MDE LookML Library. The default explore file contains an explore per default MDE type. If you are using the default configuration package in MDE, you can start analyzing your data right away. See the guide on how to use the MDE LookML Library for more details.