DATA INTEGRATED. DECISIONS ELEVATED.
Data model optimization
Overview
D&I Integrators executed a data modeling and optimization program for a leading national pharmacy retailer, spanning Snowflake and ThoughtSpot. The goal was to clean up and improve the performance of the retailer's distribution center analytics platform, which supported hundreds of liveboards used across the organization's reporting.
Over the course of the engagement, D&I audited the full pipeline from raw Snowflake assets through ThoughtSpot models to the liveboards analysts and leaders relied on daily, then rebuilt what needed rebuilding.
The Challenge
The retailer's analytics platform had grown organically over time. Multiple ThoughtSpot models were fed by a sprawling set of Snowflake tasks, tables, views, and schedules, with inconsistent naming, unclear lineage, and redundant data paths. That complexity made it harder to trust query performance, manage access, and maintain the liveboards that depended on it.
The Solution
D&I ran a structured, three phase program to audit, refine, and rebuild the platform from the ground up.
Alongside the ThoughtSpot work, D&I refactored the underlying Snowflake assets to reduce redundancy, improve performance, and create reusable outputs for the retailer's broader reporting needs. The scope also covered auditing role based access control and refresh patterns across the environment to strengthen governance and data access controls.
01
snowflake architecture
D&I inventoried the full Snowflake environment, including tasks, dynamic tables, views, schedules, ownership, and source to ThoughtSpot lineage, to build a clear picture of how data actually moved through the system.
03
performance tuning
D&I rebuilt and refined liveboards and answers on the newly optimized models, benchmarked query behavior, and tuned refresh patterns, clustering, and consumption for speed and reliability at scale.
02
ThoughtSpot Model Refinement
Using that inventory, D&I optimized the ThoughtSpot data models with clean naming conventions, reduced data redundancy, and more performant queries against dynamic tables and materialized views.
Why It Matters
By cleaning up the data model at every layer, from Snowflake schemas to ThoughtSpot logic to the liveboards themselves, the retailer's distribution center analytics platform is faster, easier to maintain, and built on assets the broader organization can reuse with confidence.