MAWDY, MAPFRE’s Assistance Unit.
MAWDY, formerly known as MAPFRE Assistance, is a global company specializing in digital distribution models. It offers services and solutions that go beyond traditional insurance offerings.
Starting from an existing BI solution, MAWDY required a technology upgrade. This new solution aims to implement a new Financial Business Intelligence system based on Atenea’s architecture and Power BI.
To achieve this, a functional and technical redesign was necessary to address performance issues while aligning with current and future needs to ensure continuity.
The Challenge: Data and Analysis
The main challenge faced by the MAWDY team was the need for a technical and functional redesign of its Financial Business Intelligence system, previously developed on on-premise infrastructure.
The project’s objective was to implement a new Financial BI system based on Atenea, MAPFRE’s corporate data platform. This would enable MAWDY to continue analyzing financial data and making informed decisions through Power BI dashboards.
Another major challenge was ensuring that the implemented ETL processes could handle large data volumes (millions of records), which vary significantly by country.
To meet these objectives, the following actions were structured:
- Design and implementation of ETL processes for automating data loading from various sources.
- Implementation of validation and quality rules during data loading.
- Creation of a DataMart for accounting data exploitation at the country level.
- Development of dashboards and reports for financial data analysis in Power BI.
- Configuration of security and user profiling.
- Power BI training.
Data Exploitation and Self-Service Models with Power BI
The implementation of MAWDY’s new Financial Business Intelligence system and the management of large data volumes, adjusted to the needs of each country, required a solution based on Atenea’s architecture. The process began with analyzing current processes and culminated in the development of self-service models in Power BI.
- Analysis of on-premise processes implemented in PowerCenter to migrate them to Python/PySpark and deploy them on AWS. MAWDY needed to migrate its on-premise Financial BI system, implemented in PowerCenter, to AWS cloud infrastructure. This involved analyzing the processes to extract business logic and applying it in the new cloud BI system, ensuring consistent results.
- ETL orchestration with Airflow. Atenea, MAPFRE’s advanced analytics platform, was integrated with Airflow’s managed service (MWAA) for orchestrating data extraction, transformation, and loading processes into the Data Warehouse.
- Creation of a DataMart in Snowflake for modeling and making financial data available. The next step involved building a star schema model composed of several fact and SCD2 dimension tables. The data was pre-processed on AWS before being loaded into Snowflake’s Data Warehouse. Views were then created in Snowflake to meet user requirements for Power BI reports.
- Design and implementation of ETL processes using Python and/or PySpark to automate data loading and processing. Given the varying data volumes across the countries where MAWDY operates, scalable ETL processes were implemented to handle different amounts of data effectively.
- Creation of a DataLake in S3 for data storage. S3 was used to store data at different stages, partitioned by country and year, enabling independent data processing and consumption for each country without disrupting operations in others.
- Development/migration of dashboards and reports in Power BI. The migration of dashboards and reports was completed in Power BI, ensuring proper data exploitation for financial analysis.
- Development of self-service models in Power BI. In the final project phase, self-service models were developed in Power BI, enabling end-users to create their own ad-hoc reports.
Efficiency in large-scale data management.
Power BI
Results in Data
- Data from 59 countries where MAWDY operates was processed.
- The new cloud-based system generates reliable, business-validated results.
- Integration of multiple data sources, including SAP, Snowflake, and files.
- Leveraging in-house advanced analytics platform accelerators (Atenea).
- Adherence to MAPFRE’s “best practices” in security and traceability.
- A new orchestration system based on Airflow was built, maximizing parallelization and reducing processing times.
- Technical foundations were established for future MAWDY projects involving SAP data processing.
- Self-service capabilities for business users were significantly improved.
- Alignment with the new account hierarchy for greater granularity and detail in reports.
- Data load times and data flow updates in the Data Warehouse were optimized.
Technologies
AWS as the cloud provider.
S3 for data storage.
Python and PySpark for ETL processes.
Airflow for process orchestration.
Snowflake for data modeling and availability.
PowerBI for data visualization and exploitation.
numbers
processed
generated
created
orchestrated