Improved Data Management and Performance Tracking for a Multinational FMCG Company

Requirement Gathering:
We discussed and noted all the requirements of the stakeholders. We understood the business model and overall functioning of various segments of the organization. About 22 detailed reports were created.

Data Model:
A scalable and less cumbersome data model with quality data and better insights was built. This model requires less maintenance and new data can be added easily without any hassle.

Data Cleaning:
Unnecessary segments were removed for better clarity. We removed redundant columns, rows, and special characters (data massaging) and renamed the parts where applicable.

Extract Transform Load (ETL):
We carried out ETL using Power Query on cloud (Dataflow, Power BI’s new feature). Data was extracted from the source (OneDrive for Business) and transformed as per the requirement (eg. changing data types of a few columns, replacing null values, etc.). Next, the data was loaded to Power BI using the Dataflows connector
Note: The ETL process was done once. The Dataflow was used to build new reports.
.

Data Storage:
Data was stored in OneDrive for Business and segregated into sub-folders as per the requirements. As a proof of concept, the endpoint of OneDrive was fed directly to the Dataflows as data input. After the approval from the stakeholders, we fed the data to an Enterprise Data Warehouse we built, which is more scalable.

Dataflow:
Since multiple Power BI reports were required, we used Dataflows. Dataflows can be integrated easily with BI tools and provides a platform to build reusable data preparation solution.

Report and Visualization:
Multiple reports were built using the same Dataflows and were customized to align with the brand color palette. Real-time visuals, graphs, charts, etc. were inculcated to understand the data better and get actionable insights.