Reference for Data Science

This guide lists the predefined objects in OCI Resource Analytics for the Data Science service. You can find information about views, entity relationships, subject areas and sample queries.

Views

This section provides information about views within OCI Resource Analytics Data Science and their columns, data types, keys, and the referred view and column names. The following views are available:

Data Science Views
Name	Description
DATA_SCIENCE_JOB_DIM_V	This view stores information about job for training models.
DATA_SCIENCE_JOB_RUN_DIM_V	This view stores information about a job run.
DATA_SCIENCE_ML_APPLICATION_DIM_V	This view stores information about a definition of an AI/ML use case.
DATA_SCIENCE_ML_APPLICATION_IMPLEMENTATION_DIM_V	This view stores information about a resource representing a solution for a AI/ML use case defined by a ML Application.
DATA_SCIENCE_ML_APPLICATION_IMPLEMENTATION_VERSION_DIM_V	This view stores information about read only fully managed snapshots of MlApplicationImplementation taken when MlApplicationImplementation was updated with a new ML Application package.
DATA_SCIENCE_ML_APPLICATION_INSTANCE_DIM_V	This view stores information about a resource representing an instance of ML Application.
DATA_SCIENCE_ML_APPLICATION_INSTANCE_VIEW_DIM_V	This view stores information about a representation of a ML Application Instance which providers use for instance observability.
DATA_SCIENCE_MODEL_DIM_V	This view stores information about models, which are mathematical representations of the relationships between data. Models are represented by their associated metadata and artifacts.
DATA_SCIENCE_MODEL_DEPLOYMENT_DIM_V	This view stores information about model deployments, which are used by data scientists to perform predictions from the model hosted on an HTTP server.
DATA_SCIENCE_MODEL_VERSION_SET_DIM_V	This view stores information about a model version set to associate different versions of machine learning models.
DATA_SCIENCE_NOTEBOOK_SESSION_DIM_V	This view stores information about notebook sessions, which are interactive coding environments for data scientists.
DATA_SCIENCE_PIPELINE_DIM_V	This view stores information about a pipeline to orchestrate and execute machine learning workflows.
DATA_SCIENCE_PIPELINE_RUN_DIM_V	This view stores information about descriptions of pipeline runs.
DATA_SCIENCE_PRIVATE_ENDPOINT_DIM_V	This view stores information about data science private endpoints.
DATA_SCIENCE_PROJECT_DIM_V	This view stores information about projects, which enable users to organize their data science work.
DATA_SCIENCE_SCHEDULE_DIM_V	This view stores information about a repeating action. Examples: Invoke a ML Pipeline Run once an hour, Call ML Job Run every night at midnight.
DATA_SCIENCE_JOB_RUN_FACT_V	Fact table for Data Science job runs.
DATA_SCIENCE_PIPELINE_RUN_FACT_V	Fact table for Data Science pipeline runs.

The suffixes in the view names designate the view type:

FACT_V: Fact
DIM_V: Dimension

The contents of each view and their relationships are listed in the following file: Data Science views.

Each tab in the Excel spreadsheet contains the details of a view.

Relationship Diagram

This section provides diagrams that define the logical relationship of a fact table with different dimension tables.

These diagrams show the relationship of the Data Science fact view with different dimension views.

DATA_SCIENCE_JOB_RUN_FACT_V

DATA_SCIENCE_PIPELINE_RUN_FACT_V

Relationships exist among dimensions. Dimensions can be joined directly to each other. These diagrams show the relationships between dimension views.

DATA_SCIENCE_ML_APPLICATION_DIM_V

DATA_SCIENCE_ML_APPLICATION_IMPLEMENTATION_DIM_V

DATA_SCIENCE_ML_APPLICATION_IMPLEMENTATION_VERSION_DIM_V

DATA_SCIENCE_ML_APPLICATION_INSTANCE_DIM_V

DATA_SCIENCE_ML_APPLICATION_INSTANCE_VIEW_DIM_V

DATA_SCIENCE_MODEL_DEPLOYMENT_DIM_V

DATA_SCIENCE_MODEL_DIM_V

DATA_SCIENCE_MODEL_VERSION_SET_DIM_V

DATA_SCIENCE_NOTEBOOK_SESSION_DIM_V

DATA_SCIENCE_PRIVATE_ENDPOINT_DIM_V

DATA_SCIENCE_SCHEDULE_DIM_V

Sample Queries

Sample queries for Data Science.

List the number of job runs associated with each job ID.

SELECT
    JOB_ID,
    COUNT(JOB_RUN_ID) AS JOB_RUN_COUNT
FROM OCIRA.DATA_SCIENCE_JOB_RUN_FACT_V F
GROUP BY JOB_ID;

List the number of pipeline runs associated with each pipeline ID.

SELECT
    PIPELINE_ID,
    COUNT(PIPELINE_RUN_ID) AS PIPELINE_RUN_COUNT
FROM OCIRA.DATA_SCIENCE_PIPELINE_RUN_FACT_V F
GROUP BY PIPELINE_ID;

Data Lineage

The Customer Experience Semantic Model Lineage spreadsheet and Metric Calculation Logic spreadsheet for Data Science provide an end-to-end data lineage summary report for physical and logical relationships in your data.

For more information, see Data Lineage.

Subject Areas

This section provides information on the subject areas with data you maintain in Data Science. These subject areas, with their corresponding data, are available for you to use when creating and editing analyses and reports. The information for each subject area includes:

Description of the subject area.
Business questions that can be answered by data in the subject area, with a link to more detailed information about each business question.
Job-specific groups and duty roles that can be used to secure access to the subject area, with a link to more detailed information about each job role and duty role.
Primary navigation to the work area that's represented by the subject area.
Time reporting considerations in using the subject area, such as whether the subject area reports historical data or only the current data. Historical reporting refers to reporting on historical transactional data in a subject area. With a few exceptions, all dimensional data are current as of the primary transaction dates or system date.
The lowest grain of transactional data in a subject area. The lowest transactional data grain decides how data are joined in a report.
Special considerations, tips, and things to look out for in using the subject area to create analyses and reports.

The subject area is:

Data Science

Useful Resources

Oracle Cloud Infrastructure Documentation / API Reference