Jason Hull

The project was started after the company spun off from its parent company as a standalone entity. Initially, a SQL server farm stood up to duplicate warehousing, analysis, and reporting services.

After the spin-off, a project began modernizing the data warehousing to use the cloud. Snowflake, Fivetran, and dbt were the selected tools.

My role was to understand, catalog, and reverse engineer the existing SSRS reporting based on T-SQL queries to the SQL instance and MDX queries to the SSAS server. Once understood, I then had to build the model in dbt Cloud and test it to ensure the accuracy of the new structure.

In dbt, I used dbt Cloud with a combination of snapshotting for type 1/2 dimensions, traditional fact structures, and a big-wide-table methodology to ease the report building in Tableau, all using the Fishtown syntax. Additional dbt functionality includes utilities from dbt-labs for common macros, custom macros that fit the project's purpose, and multiple pipeline jobs to break out the data warehouse build.

This part of the data mart was made up of multiple dimensions and a single fact table that read from multiple billing system sources to identify what the value (incoming revenue) of a customer is and trends that data over time.

This query was pulled from an existing set of call records and made into a support center. Using database analytic functions to see if the same account number called in on the same day, it labeled calls and resolved them if they were the last call made or only a call from an account number that remained unresolved. A report on top of this was then used to display the number of unresolved calls and a first call resolution rate.

This example MDX query is handwritten to pull in all the key metrics from an SSAS cube to be displayed on a dashboard. The date parameter is dynamic based on user selection, and some members are derived based on other members.

Developed an Oracle PL/SQL procedure that accepts input parameters to dynamically delete records from the table passed in. I also tuned and maintained it over time.

Developed a query embedded in the view that was timing out after hours when submitted to a database via an application. I moved it into a daily materialized view refresh, saving time and improving stability.

Developed a looping PL/SQL procedure used to read update statements stored in a table with a configurable commit level.

I served as the lead developer during a migration project away from SSIS and IBM Cognos Data Manager as ETL platforms. My work included installing and administering Informatica PowerCenter.
DELIVERABLES
• Install and configure the PowerCenter engine on a Linux host.
• Install the client on all developers' desktops.
• Define and build standards.
• Identify reusable items to convert into maplets or workouts.
• Design, code, and deploy all core mapping and workflow components.
• Review of junior members' code.
• Ensure standards were met for the use of lookups, expressions, and target loads.
• Lead the discussion on when to use ETL vs. ELT approach.
• Integrate the enterprise scheduler.