New ways to optimize shared datasets in the Power BI Service
summary
A demo shows a concrete workflow that finds and removes unused columns and measures from a shared Power BI dataset, using Measure Killer locally (PBIX + downloaded reports) and Tabular Editor 3 via the XMLA endpoint to edit M code online. The analysis checks visuals, conditional formatting and filters, reports a surprising 50% unused (941 MB → 427 MB in the demo), supports restore/backups and Excel export, and finishes with a dataset the tool marks 100% used while noting fringe limitations like copied pages and table-count edge cases.
AI-written summary · about
chapters (11)
- 0:00Shared datasets in Power BI
- 3:21Optimizing shared datasets
- 7:21Live demo: shared dataset model
- 10:08Analyzing unused elements
- 22:39Cleaning measures and columns
- 27:39Removing unused Power Query columns
- 32:13Saving optimized model
- 40:24Online optimization without Desktop
- 44:45Tabular Editor cleanup
- 47:16Finalizing results
- 49:02Q&A
related talks
18mEmpowering Business Through Data
51mPower BI Composite and Hybrid Models
45mRebooting your SME Data Analysis Strategy
54mExtending PowerPivot into a Scalable Component for Enterprise BI
02h 12mOptimizing your Power Pivot models using Power Query
54mData Visualization with Power View and the Tabular Model
48mDon't be scared with Hybrid, they are only tables…
01h 04mBest Practices for Collecting and Analyzing Data with Power Query and Excel
