Use MaxCompute Notebook

Step 1: Create an instance template

You can create an instance template on the Notebook page. You can then reference this template when you create Notebook instances.

1. Go to the Notebook list page. On the Instance Template tab, click Create Instance Template.

2. In the Create Instance Template dialog box, configure the following parameters.

   Parameter	Description
   Instance Template Name	The name of the Notebook instance template to be created.
   Description	A description of the instance template.
   Automatic Release Settings	The time when the instance is automatically released. No: The instance is not automatically released. Yes: The instance is automatically released after a specified number of hours.
   Select Compute Engine	The MaxFrame software development kit (SDK) is built-in and ready to use.

3. Click OK to create the instance template.

Step 2: Create a Notebook instance

1. Log on to the MaxCompute console. In the navigation pane, choose Workspace > Notebook to go to the Notebook list page.

2. On the Instance Management tab, click Create Instance and configure the parameters.

   Parameter Name	Description
   Instance Name	The name of the Notebook instance to be created.
   Description	A description of the instance.
   Associated Project	This lets you perform subsequent data operations on associated projects without an AK/SK.
   Instance Creation Method	You can create an instance from scratch or from an instance template.
   Select Instance Template	This parameter is required when you set Instance Creation Method to From Instance Template. Select an existing instance template or create one. For more information, see Step 1: Create an instance template.
   Automatic Release Settings	This parameter is required when you set Instance Creation Method to From Scratch. No: The instance is not automatically released. Yes: Specify the time after which the instance is automatically released.
   Computing Resources	Select a quota group. Note Currently, only pay-as-you-go quota groups are supported.
   Storage Configuration	Select an existing data storage. You can mount a user-created NAS file system for script file persistence. You can also click Create Data Storage to create a new one. For more information, see Appendix: Create a data storage.
   Compute Engine	The MaxFrame SDK is built-in and ready to use.
   Is this instance shared?	Visible to Tenant: The instance is visible to all users within the tenant. Visible to Me Only: The instance is visible only to you and the administrator.

3. Click OK. The instance is ready to use when its Status changes to Running.

   Note

   After the instance is created, you can also click > Auto-release Settings in the Actions column of the target instance to change the auto-release configuration.

Step 3: Develop in Notebook

To help you quickly get started with development, MaxCompute Notebook provides a demo script. The script is based on MaxFrame and shows how to perform distributed processing with Pandas. It covers data preparation, data analytics, data exploration, and distributed data processing. You can click product_sales_demo_nb.ipynb to download the script.

1. Go to the Notebook instance page and upload the demo script.

   In the Actions column of the target instance, click Enter. On the instance page, click the icon on the left to upload the demo script.

2. Enter the project information, execute the script file, and generate the visualization.

   1. In the pane on the left, double-click the product_sales_demo_nb.ipynb script to open it. Follow the instructions in the script to set the PROJECT_NAME. You can use the project associated with the Notebook instance or another project.

      • Use the associated project:

        In the script file, find the Create ODPS object code block in the 2.2 Prepare data section. Replace project=PROJECT_NAME with project=os.getenv('ODPS_PROJECT_NAME'). In this case, you do not need to specify PROJECT_NAME in the 2.1 Prepare project section.

      • Use another project:

        Set PROJECT_NAME to the name of the MaxCompute project that you want to use for computation.

   2. At the top of the script file, click the icon to run all code blocks. In the Restart Kernel? dialog box, click Restart. The script is then executed. The execution is successful when the circle icon in the upper-right corner of the page changes to an empty circle and no errors are reported in the code.

   3. View the chart generated by Matplotlib to visualize the data mining and analysis results.

Step 4 (Optional): Release the Notebook instance

After development is complete, go to the Instance Management tab on the Notebook page. In the Actions column of the target instance, click Stop and then click Delete to release the Notebook instance.