GreenGrid Solutions - The "Smart Pricing" Pilot (Microsoft Fabric)

 Company Background

GreenGrid Solutions is an energy service provider piloting a smart electricity billing model for residential customers. The company is shifting from flat-rate billing to dynamic, peak-hour pricing to reduce grid strain and promote efficient energy consumption.

Project Overview
The pilot faced operational challenges due to inconsistent meter readings, manual billing logic, and data privacy risks. Raw meter data contained invalid values, pricing calculations lacked standardization, and customer email IDs were exposed in downstream datasets. The project delivered an end-to-end data engineering solution using Microsoft Fabric, covering ingestion, cleansing, enrichment, privacy protection, and billing calculations within a Lakehouse architecture.

Problem Statement
Meter data included invalid and unrealistic kWh values. Billing calculations were manual and error-prone. Customer email IDs were exposed in analytical datasets. There was no separation between valid and invalid billing records, and the solution lacked scalability for future smart meter expansion.

Project Objectives
Ingest customer, pricing, and meter data into a centralized Lakehouse. Validate meter readings and quarantine anomalous data without disrupting pipelines. Enrich usage data with customer plans and pricing rules. Apply peak-hour pricing logic. Protect customer email IDs using hashing techniques. Deliver a clean, analytics-ready Silver billing table.

Solution Overview
The solution follows a Medallion Architecture implemented in Microsoft Fabric.

Bronze Layer (Raw Ingestion)
Customer PII data, pricing plans, and smart meter readings are ingested as-is into the Lakehouse.

Transformation Layer
Data quality validation is applied, usage data is enriched through joins, customer email IDs are protected using hashing, and business rules including peak-hour pricing are enforced.

Silver Layer
A clean, privacy-safe billing table is produced along with a dedicated quarantine table for rejected records.

Data Ingestion
A Microsoft Fabric data pipeline was built with three copy activities for customer, pricing, and meter datasets. All source files were reliably and repeatedly ingested into the Bronze layer before transformation.

Business Outcomes
Billing inaccuracies caused by faulty sensor data were eliminated. Peak-hour pricing calculations were fully automated. Full PII compliance was achieved for analytics consumption. Manual Excel-based workflows were replaced with scalable pipelines, and a reusable smart meter billing framework was established.