Student Capstone Project Spotlight

James Freeman

Program: Computer Systems Technician – Information Technology Infrastructure and Services
Capstone Project: Barback

Artificial Intelligence (AI) and Machine Learning (ML) are growing industries demanding highly skilled graduates. Dr. Jonathan Kim, in collaboration with FAST faculty members, staff, members of the Program Advisory Committee (PAC) and industry partners, have recently developed a new graduate certificate in AI and ML. Designed for students with experience in computer programming, software development or computer science, students will learn the fundamentals to apply AI models and ML algorithms to a broad range of practical applications including natural language processing, image processing, computer vision and for business decision making.

While this graduate certificate takes flight, we look back at a 2022 student ML capstone project that showcases this growing industry.

Tell us about your project?

Today more than ever, it is important to carry only the product that you need to keep cash freed up for use in other areas of the business. To do so you need to find the balance between carrying too little product, while simultaneously being wary of buying too much. It can be difficult to manage inventory levels due to the vast number of drinks that can possibly be ordered. A single drink can have as little as 2 and up to 10 or more different ingredients. Not only that, but bar sales can often be affected greatly by things like seasonality, location, or weather as well. I set out to create a system that would take all these factors into consideration and help bar owners and with the decision making involved in managing their bar’s inventory.

I have noticed over the years that one thing each bar manager had in common was the way they decided how much product to order; They simply looked at how much product was sold during this time last year. While that offers some information it is not very reliable. Maybe the bar is relatively new, so those numbers don’t exist, or maybe the bar has been open for two years but only recently started to see a significant increase in sales; Those numbers are no longer relevant.

So, I set out to create a system that would offer as much guidance when making this decision. I implemented a way to use machine learning along with Excel and Python that will predict how much product a bar will sell in each week. The goal of this project is to provide decision support to bar managers to help them optimize their ordering strategy.

Machine Learning (ML) is a way to collect data to answer questions. In this case the data we are collecting is previous liquor sales, and the question we are answering is how much of an item will a bar sell next week? There are so many other industries, like medicine or improving autonomous vehicles, that have implemented ML in some form or another with great success. If it has worked so well for these other businesses, why can’t it work just as well for a bar? So, I created a system that puts ML to work to improve the guess work involved in determining how much product is required in each week. The goal of this project is to provide decision support to bar managers to help them optimize their ordering strategy.

What are you most proud of?

I am proud that I was able to create a unique project that was built from the ground up with the guidance of Timothy Wong. This project required a lot of resourcefulness and problem solving. In the end I was able to combine my work experience with my studies to create a system that has a real-world application.