How Open Source Big Data Solution Boosted Giant Retailers' Revenue By 32%

Client Challenges

• Though the client has a wide network of physical stores, they rightly guessed the importance of digitalization and started out with an online store
• To devise the Omni channel model of business, the client created a web of multiple online and physical channels to let customers interact and buy from them
• Due to their inexperience in comprehending digital workflows and managing online data chunks, they failed to aggregate and analyze data from multiple channels and multiple data formats including structured and unstructured files
• It became almost impossible for the client to contextualize consumer behavior, as collected consumer data wasn’t just discreet but non-usable
• Consequently, the client was not able to create a behavior-based profile of all the visitors on the site and trigger a personalized recommendation of products
• It significantly damaged the sales of the online wing of the client’s business

Technology Stacks

• Hadoop
• Mahout
• Python

Our Solutions

• A thorough examination of the digital initiatives of the client helped us to identify the data analysis crisis, the real pain point of the client
• We began by observing and recording the total amount of data that client’s digital avenues used to generate
• Based on it, we scientifically predicted its year on year growth to make the proposed solution future proof
• Looking at the huge amount of structured and unstructured data that the client had to handle, there was a need for the visualization of unstructured data or the ways to link structured data to unstructured data
• Considering that too-real need, we implemented an open source Big Data solution (Hive/Pig) to process massive amounts of structured and unstructured data
• To securely store the processed data, we proposed the implementation of Hortonworks Hadoop
• To make data actionable through analytics, we created a collaborative map by reducing filtering implementation in Hortonworks