Skip to main content

Implement AXS analytics

This guide explains how to implement and manage the AXS analytics workflow, including key dependencies, challenges, and step-by-step instructions for creating product analytics dashboards.

Overview of the Process​

The AXS analytics workflow follows the ETL paradigm (Extract, Transform, Load). The process involves:

AXS Analytics Workflow

  1. Defining and implementing analytics events in the app.
  2. Ingesting events from Azure Application Insights into Snowflake.
  3. Transforming data in Snowflake using custom SQL views.
  4. Building a semantic model in Power BI Desktop.
  5. Creating dashboards in Power BI Web.

Step-by-Step Guide​

Step 1: Define and Implement Analytics Events​

Step 2: Request Data Ingestion​

  • Submit a Data Access Request via the Galileo Portal.
  • If the data source is already configured (e.g., AXS - Application Logs), request access directly. Otherwise, request a new ingestion pipeline.

Step 3: Create Snowflake Views​

  • Ensure you have the necessary permissions in Snowflake.
  • Use the Worksheets tab to create a SQL view for transforming raw data.
  • Example SQL view:
CREATE OR REPLACE VIEW AXS_DATA_PRODUCTS.APPLICATION_LOGS.SIAB_REVIEWER_TEST_VIEW AS (
        SELECT
                REGEXP_REPLACE(properties:contactId, '^00', '_', 2) AS sap_user_id,
                DATE_TRUNC('day', TO_DATE(timegenerated)) AS login_date,
                'Q' || QUARTER(TO_DATE(timegenerated)) || ' ' || YEAR(TO_DATE(timegenerated)) AS q,
                WEEKOFYEAR(TO_DATE(timegenerated)) AS w,
                properties:organizationId AS sap_customer_id,
                properties:organizationCountry AS customer_country_code,
                NAME, TIMEGENERATED, USERID, SESSIONID, CLIENTBROWSER
        FROM AXS.APPLICATION_LOGS.USE_APPEVENTS
        WHERE REGEXP_LIKE(NAME, '.*REVIEWER_.*')
        GROUP BY 1
);

Step 4: Build Semantic Model in Power BI Desktop​

  • Use Power BI Desktop (Windows-only) to connect to Snowflake.
  • Input the server (mg41896.eu-west-1.snowflakecomputing.com) and warehouse (COMPUTE_WH) credentials.
  • Select the Snowflake view and create the semantic model.

Step 5: Create Dashboards in Power BI Web​

  • Use the Power BI Web app to design dashboards based on the semantic model.
  • Example lineage and report creation:

Dependencies​

  • Platforms: Azure Application Insights, Snowflake, Power BI.
  • Teams: Engineering, Data Analysts, IT/Platform, Galileo team.

Challenges​

  • Lack of documentation and cross-domain expertise.
  • Power BI Desktop requires a Windows machine.
  • Ownership of dashboards is unclear.

References​