Business Analytics and Intelligence

This is one of the first assignments we worked on in the course. An exploratory data analysis (EDA) is a method analysts use to get a better feel for the data. There isn't necessarily a step-by-step process to follow, one just uses a range of commands to inquire into the basic characteristics of the data before attempting to solve any problems.

For this assignment, we were given a data file and tasked with finding which variable (or variables) are the best predictor(s) for sales for a local grocery outlet.

In conjunction with the previous assignment, this document is a report that would be sent to the management of the grocery outlet with actionable feedback concerning what we found with the multiple regression analysis. For reference, we were asked not to use any analytical terminology or code, because we should write as though we were reporting to someone who knows little about the field of analytics and merely wants to improve the profitability of their grocery chain.

I thoroughly enjoyed working on this assignment. For this project, we were given data on Likert-Scale responses asking consumers about their purchasing and leisure preferences. First, we had to guess at the optimal number of clusters, then we went back and attempted to justify our guess using k-means clustering methods to distill groups. Before we justified our clustering guesses, however, we put forth ideas on how we would market to each group given their purchasing and leisure preferences. This was quite enjoyable because it was the point in the semester where our homework assignments were becoming less clear-cut and more nebulous.

As I'm told, this assignment probably needs little introduction because it is quite standard practice in the field when it comes to determining consumers' creditworthiness when applying for a loan. Here I used decision trees to find the best discriminators when trying to predict whether a potential client will default on a loan.

This was the first half of our final exam in the course. For this project, I was given a data set which detailed numerous characteristics for various red wines. One of the columns in the data set let us know if the wine scored a 'good' rating or not in a widely-published wine-tasting magazine. At the completion of my analysis, I provided a recommendation for how to engineer the 'best-tasting wine' according to the magazine's standards.

For the 2nd half of our exam, we were given a data file that detailed various clickstream data and purchasing behaviors of online shoppers. We were asked to determine the optimal number of clusters, identify significant distinctions across the groups and develop a tailored marketing strategy for each group. I truly revel in indefinite assignments like this one because it offers me an opportunity to showcase my creativity and problem-solving skills.