Our client, a global retailer, was facing multiple challenges related to their legacy system for reporting and visualizing Market Basket Analysis (MBA) metrics. While the cost and effort of maintenance is high most of the current requirements of the business users are not met by the legacy system. Further, attempting to change the legacy system to incorporate new requirements is time consuming and not scalable.
Resolve Tech team worked as part of the client’s Business Intelligence and Data Analytics team to replace this legacy system with contemporary technology stack and scalable deisng & architecture for data capture, reports, and visualization. The solution was implemented with key technologies such as Azure Data Lake, Apache Spark, and PowerBI. Besides, the technology upgrade, the team implemented new requirements of various business groups.
Raw data from different stores across all the products and sales activities need to be cleansed (for standardized codes and acronyms, removing duplicates, filling default values etc.) and specific transformation rules many need to applied for specific data sets.
Data for more products and stores are considered for better data quality across different time ranges (Year, week, month, and year-to-date etc.)
Designed and implemented highly scalable data stores & pipelines leveraging open-source technologies and streamlined data pipelines that are highly optimized for large data sets thus ensuring high-throughput and low-latency operations.
Ensure that the calculations and rules applied in the legacy reports are transferred accurately into the new reports while also making sure new data & visualization requirements from different business groups are implemented.
Aggregation capabilities at different organization hierarchy levels and across different attributes of product hierarchy (Attributes such as item-name, business-unit etc.)
Provide data visualization in different forms from tabular reports to charts, bar graphs, stacked bar graphs etc. including the facility for the user to drill-down to multiple levels and download the underlying data to conduct off-line custom analysis.