Compare Tableau live vs extract and filters

Read the full interview experience this question came from →

Quick Overview

This question evaluates competency in BI data architecture and large-scale dashboard performance, covering extract vs live connections, refresh and incremental strategies, query pushdown versus in-tool processing, row-level security, filter order-of-operations, performance tuning, and platform migration mapping within the Analytics & Experimentation domain for a Data Scientist role. It is commonly asked to assess trade-off reasoning across data freshness, storage, query performance and security for a 100M-row fact table, combining practical application-level decisions (implementation and tuning) with conceptual understanding (filter ordering, calculations, and migration pitfalls) when moving between BI platforms.

Compare Tableau live vs extract and filters

Company: Amazon

Role: Data Scientist

Category: Analytics & Experimentation

Difficulty: hard

Interview Round: Onsite

In Tableau, compare Live vs Extract connections for a dashboard over a 100M-row fact table. Discuss refresh cadence, query pushdown, extract size, incremental refresh, row-level security, and when each is preferable. Explain the order of operations and how it affects: - Data source filters, context filters, dimension/measure filters, TOP/N, table calculations, and FIXED/INCLUDE/EXCLUDE LODs. - Performance tuning: selecting dimensions for context, using extract filters, avoiding expensive table calcs, and denormalizing vs joins/relationships. Map these ideas to Amazon QuickSight analogs (SPICE vs direct query, row-level security, filters) and note one pitfall when migrating a Tableau workbook to QuickSight.

Overview: This question evaluates competency in BI data architecture and large-scale dashboard performance, covering extract vs live connections, refresh and incremental strategies, query pushdown versus in-tool processing, row-level security, filter order-of-operations, performance tuning, and platform migration mapping within the Analytics & Experimentation domain for a Data Scientist role. It is commonly asked to assess trade-off reasoning across data freshness, storage, query performance and security for a 100M-row fact table, combining practical application-level decisions (implementation and tuning) with conceptual understanding (filter ordering, calculations, and migration pitfalls) when moving between BI platforms.

Read the full Amazon Data Scientist interview experience this question came from

|Home/Analytics & Experimentation/Amazon
Amazon logo
Amazon
Oct 13, 2025
hardData ScientistOnsiteAnalytics & Experimentation
7
0

Scenario

You are building an interactive dashboard over a 100M-row fact table. Compare Tableau connection options and performance behaviors for this scale, and map the concepts to Amazon QuickSight for a potential migration.

Part A — Tableau: Live vs Extract for 100M rows

Discuss the trade-offs between Live and Extract connections, specifically addressing:

  1. Refresh cadence and data freshness.
  2. Query pushdown to the database vs Tableau Hyper.
  3. Extract size and storage considerations.
  4. Incremental refresh design (late-arriving data, updates vs inserts).
  5. Row-level security (RLS) enforcement and data leakage risks.
  6. When each option is preferable for a 100M-row fact table.

Part B — Tableau Order of Operations and Effects

Explain Tableau's order of operations and how it affects the following:

  • Data source filters
  • Context filters
  • Dimension filters
  • Measure filters
  • TOP/N filters
  • Table calculations
  • LOD expressions: FIXED, INCLUDE, EXCLUDE

Part C — Tableau Performance Tuning Tactics

Explain how to tune performance for this dashboard by:

  • Selecting high-selectivity dimensions for context filters.
  • Using extract filters and pre-aggregation.
  • Avoiding expensive table calculations when possible.
  • Choosing denormalization vs joins/relationships appropriately.

Part D — Mapping to Amazon QuickSight

Map the above ideas to Amazon QuickSight analogs and note one migration pitfall:

  • SPICE vs Direct Query (including refresh/incremental, pushdown, size).
  • Row-level security approaches.
  • Filters and calculation analogs.
  • One common pitfall when migrating a Tableau workbook to QuickSight.
Loading comments...

Browse More Questions

More Analytics & Experimentation•More Amazon•More Data Scientist•Amazon Data Scientist•Amazon Analytics & Experimentation•Data Scientist Analytics & Experimentation