Through June and July of 2022, I got to be a member of the Digital Footprints Lab at the University of Bristol. I worked primarily with Dr. Neo Poon and Dr. Anya Skatova, and the goals of this internship were to better understand the uses of entropy measures in consumer research as well as get a better idea of what working with data looks like in an academic setting.
My tasks were outlined to best fit my time in Bristol, and I got to be in the office at Oakfield House at the University of Bristol Medical School with Neo and Anya multiple days a week which was a great part of my internship. I used the programming language R to accomplish my tasks of simulating, cleaning, and visualizing data. More specifically, I was tasked with better understanding the different measures of entropy we can use in consumer research and then building a relatively simple simulated shopping environment to collect sample data and build plots from. The goal of this process was to better understand each entropy metric (unnormalized and types of normalized entropy) and why certain ones may be better than others, as it is the case that there is no consensus best entropy metric.
After dealing with simulated data, I moved to more real world data from the dunnhumby database for consumer research. Their sample transaction data has over 31 million individual transactions with around 500 thousand unique customers, and with 25 variables measured for each transaction. Although working with this data takes a lot of computational power, some valuable work can be done with it and I was glad I got to transition to real world data at the end of my internship.
My time in Bristol working with the Digital Footprints Lab was extremely rewarding and valuable. The group was nothing but welcoming to me and showed me a glimpse into the great work they all do. My work with Neo and Anya has definitely sparked an interest in data analysis in my career going forward, and I am delighted I got the opportunity and am excited for any in the future, should they arise.