Fabric Finance Reconciliation Demo

A synthetic Microsoft Fabric demo showing how a manual month-end GL-to-subledger reconciliation can move into a controlled Lakehouse and Power BI workflow.

Demo flow

flowchart LR
    A[Source CSV files] --> B[Lakehouse Files: Bronze raw evidence]
    B --> C{Processing path}
    C --> D[Option A: Dataflow Gen2]
    C --> E[Option B: Fabric notebooks]
    D --> F[Silver standardized tables]
    E --> F
    F --> G[Gold reconciliation outputs]
    G --> H[Gold dimension tables]
    G --> I[Validation checks]
    H --> J[Power BI semantic model]
    G --> J
    J --> K[Thin Power BI report]
    K --> L[Finance_Reconciliation_Report.pbix]

Loading

What this demo does

Manual Excel reconciliation
→ Raw files stored in Fabric
→ Silver standardized tables
→ Gold reconciliation outputs and dimensions
→ Power BI semantic model
→ Thin Power BI report

Who this is for

Finance, FP&A, accounting, audit, and analytics users who want a practical example of moving a spreadsheet reconciliation process into Microsoft Fabric.

Design principle

The repo avoids creating tables unless the table has a clear business use.

Bronze = raw CSV files in Lakehouse Files
Silver = standardized GL and subledger tables
Gold = reconciliation outputs and semantic model dimensions

Repo map

Folder	Purpose
00_source_data/	Synthetic CSV source files and expected validation outputs.
01_option_a_dataflow_gen2_pqt/	Low-code Dataflow Gen2 implementation.
02_option_b_notebooks/	Code-first notebook implementation.
03_power_bi_reporting/	One Power BI instruction document plus required report assets.
04_validation/	Expected results and validation checklist.
05_runbook/	Demo runbook and validation report.
docs/	Data dictionary, schema alignment, and process review.

Start here

Open:

START_HERE.md

Output tables

Silver:

silver_gl_clean
silver_subledger_clean

Gold facts/outputs:

gold_reconciliation_detail
gold_reconciliation_summary
gold_reconciliation_exceptions
gold_file_audit_log

Gold dimensions:

dim_reconciliation_status
dim_fiscal_period
dim_account
dim_department
dim_project
dim_vendor

Expected reconciliation counts

ReconciliationStatus	ItemCount
Matched	90
Amount Mismatch	8
Timing Difference	6
Missing from Subledger	5
Missing from GL	7
Duplicate in Subledger	2

Power BI

Power BI should be built semantic-model-first:

Lakehouse tables
→ Finance_Reconciliation_Semantic_Model
→ Finance_Reconciliation_Report.pbix

Use the single Power BI instruction document:

03_power_bi_reporting/Power_BI_Report_Build_Guide.md

Disclaimer

This repository contains a synthetic finance reconciliation demo. The data is fictional and does not represent any real company, client, employee, vendor, or financial system.