Sales and Customer Analytics Tool

This tool was a program I built for my old boss at RJ Corp India’s Nike and Rookie division as part of my IB Computer Science Internal Assessment project. It was a desktop application built with Qt and Python, and aimed to automate and assist parts of the workforce of the company that I had identified during my time interning there.

This program had 2 parts to it. The first was a sales analytics tool, built for the visual merchandising team of the company, which performed association mining on their sales data to identify commonly purchased together items in their brick and mortar stores in shopping centres across the Delhi, Gurgaon, and Noida regions of India. I implemented the frequent-pattern growth algorithm and pretrained it on over 2 million sales from 2019-2021. This tool was used by the company for optimizing product locations in store, and in their recommendation algorithm for their online storefront when it was first released.

The second part of the program was a customer feedback analytics tool. When I was interning I learned that the customer feedback the company received was underutilized due to the sheer volume of it, and the lack of time for anyone to address the concerns and highlight the compliments that it contained. As such, I proposed and built a recurrent neural network (RNN) for classifying the customer feedback into distinct categories. I manually labelled a portion of the dataset according to the most overarching themes I could see in the feedback, so that the company could understand the biggest trends in their customer feedback and the magnitude of those trends.