Project: Creative Apps - Power BI Fixer

[KornAlexander]

Track

Creative Apps (GitHub Copilot)

Project Name

Power BI Fixer

GitHub Username

KornAlexander

Repository URL

https://github.com/KornAlexander/pbi_fixer

Project Description

The Power BI Report Fixer is a free, open-source tool that lets you assess, standardize, and fix Power BI reports and their semantic models from a single function call in a Microsoft Fabric Notebook.

One line of Python — pbi_fixer() — launches an interactive widget UI that combines 11 fixers across two categories:

Report Fixers (5): Replace pie charts with bar charts, clean up bar and column chart formatting (remove axis titles, add data labels, remove gridlines), upgrade default page sizes to Full HD, and hide visual-level filters from end users. These work through the PBIR format, reading and rewriting visual.json and page.json files directly.

Semantic Model Fixers (6): Set DiscourageImplicitMeasures to True, add a calculated calendar table with hierarchies and display folders, add an empty measure table, add a Last Refresh table with M partition, and create Units and Time Intelligence calculation groups. These work through the XMLA endpoint via the Tabular Object Model.

The tool supports three modes: Fix (apply changes), Scan (assess without changing anything), and Scan + Fix (scan first, then apply). Every fixer is idempotent — it checks whether its target change already exists and skips if so.

Built entirely with Python and ipywidgets on top of Semantic Link Labs. GitHub Copilot was used extensively during development for code generation, refactoring, and documentation. The tool requires no installs, no downloads, and no license costs — just one line of Python in a Fabric Notebook.

Demo Video or Screenshots

https://actionablereporting.com/2026/02/22/power-bi-fixer-one-line-to-fix-them-all/
https://www.linkedin.com/posts/alexanderkorn_fabricfomoisreal-ibcs-powerbi-activity-7431593874826874880-T-ER?utm_source=share&utm_medium=member_desktop&rcm=ACoAABEpJQUB4gUMhzlO-gPscjJoOX2bMgrL2yk

Primary Programming Language

Python

Key Technologies Used

GitHub Copilot

Python
ipywidgets (Python library for creating interactive HTML widgets)
Microsoft Fabric Notebooks
Semantic Link Labs (open-source library)
PBIR format (Power BI Enhanced Report Format)
XMLA endpoint
Tabular Object Model (TOM)

Submission Type

Individual

Team Members

No response

Submission Requirements

• My project meets the track-specific challenge requirements
• My repository includes a comprehensive README.md with setup instructions
• My code does not contain hardcoded API keys or secrets
• I have included demo materials (video or screenshots)
• My project is my own work with proper attribution for any third-party code
• I agree to the Code of Conduct
• I have read and agree to the Disclaimer
• My submission does NOT contain any confidential, proprietary, or sensitive information
• I confirm I have the rights to submit this content and grant the necessary licenses

Quick Setup Summary

1. Open a Microsoft Fabric Notebook
2. In the first cell, run:
   %pip install git+https://github.com/KornAlexander/semantic-link-labs.git
   from sempy_labs import pbi_fixer
3. In the next cell, run: pbi_fixer()
4. An interactive UI appears — select your workspace, report, fixers, and click Run.

Full details in the repository README.

Technical Highlights

What I am most proud of:

• The entire tool is a single function call — pbi_fixer() — that opens an interactive ipywidgets UI inside a Fabric Notebook. No CLI, no config files, no separate app. Just run one line and you get a full-featured fixer application.
• Each of the 11 fixers is self-contained and idempotent. They check whether the target change already exists before applying it (e.g., "does a calendar table with DataCategory=Time already exist?"), making them safe to run repeatedly.
• The report fixers work directly on the PBIR format — reading and rewriting JSON files inside the report definition — while the semantic model fixers use the XMLA endpoint via TOM. Combining both layers in one UI is unusual and powerful.
• Scan mode lets you preview every change before committing — making it safe for production use across large report estates.
• GitHub Copilot was instrumental throughout the development process, from generating the DAX expressions for calculation groups to building the ipywidgets layout and writing the fixer logic.

Challenges & Learnings

• Working with the PBIR format required reverse-engineering the JSON structure of visual.json and page.json files. Documentation is sparse, so understanding the schema involved inspecting many reports manually.
• The XMLA write operation is irreversible (the .pbix can no longer embed data), which is why I built the confirmation checkbox and scan mode — to give users full control before any write happens.
• Calculation groups interact with all measures in a model, so designing the Units calc group to skip percentage and ratio measures required careful testing across different model shapes.
• Building a polished ipywidgets UI that works reliably inside Fabric Notebooks (which have their own rendering constraints) involved iterating on layout, styling, and event handling.

Contact Information

alkorn@microsoft.com

Country/Region

Germany