Maximo Predict

Prerequisite

1. Ensure that you have completed the Maximo Monitor hands-on Lab
2. Ensure that you have completed the Maximo Health hands-on lab

Note: Understanding & Availability of sensor data sets in Monitor application and asset data in Manage application is important.

3. You will need to download the Predict_Envs.json available here.
4. You will need the download db2_certificate.pem available here.
5. You will need the download ca_public_cert.pem available here.
6. Create your own user in CP4D to run predict models.

Setup Manage Application

1. Open Manage application.
2. Go to the Assets application and filter the asset records created in Maximo Health lab exercise. e.g XX_ASSET%
3. Change the status of any two to three asset to DECOMMISSIONED e.g you can pick XX_ASSET2 and XX_ASSET4 and change the status from Active to DECOMMISSIONED.
4. Make sure you populate Installation Date, Expected life in years and Estimated EOL fields populated for each asset, without this Predict models will not execute.

Setup Predict Application

1. Open the Predict application from Suite Navigator.

2. To create a Predict Group. Click on left menu and select -> Predict Grouping

3. Click on blue Create group + button to create a new group.
4. Provide the name and description as: XX_predictscores
5. Select the query you created in Manage application e.g xx_asset

6. Click on Create button.
7. Verify the group has been created and note down the value for Group Id column. Here in below screenshot it is: 1005

Setup Cloud Pak for Data

1. Use the CP4D url and credential supplied to access the instance

Note: This will be supplied by instructors during the workshop

2. Click on left hamburger menu and select All Projects

3. Click on blue button New Project + to create a new project.

4. Choose Create an empty Project.
5. Enter project name and description as XXPREDICTSCORE

6. Click on Create button

7. Select the Assets tab and click on Drop data files section and select the below files that you saved from earlier.
   • Predict_Envs.json
   • db2_certificate.pem
   • ca_public_cert.pem

8. Select Assets tab and click on New Assets + button

9. Select Code Editor and then choose Jupyter notebook editor

10. Click on From File tab and drag and drop the PMI – End of life curve.ipynb. You can download the notebook from here

11. Click on the Create button. It will then open the PMI – End of life curve notebook.

12. Change the asset_group_id variable to match the one noted from the predict application earlier. In this case it is 1005.

13. Select first cell and click on the Run button. It executes each cell one after other.

Note: A cell will be done executing the code within when the In [*] value to the left of the cell is filled with a number like In [1].

14. Verify the output for Train the model instance. It should display "Finished execution of End of Life Curve Asset Group Pipeline." At the end.

15. Execute each cell until you reach the Register the trained model instance cell.
16. The final outcome should display the message: Registration was successful. New model ID = 20BB65D9-BA25-4173-95EC-A9E7E58DA5C7

Verify the scores

1. Go to Predict Application and open the Predict group created earlier.
2. Click on the Trained Instance model link.

3. Choose options as below:

Active : ON
Run every : 1 Day
Starting At : 9 AM
Date : Enter future date
text
Copy code to clipboard

4. Click on the Save button.
5. Open any asset in the list and verify that End of life curve is populated like below:

Congratulations! You have successfully completed the Maximo Predict Lab.