Country Fried Creative (CFC), a local marketing agency based in Peachtree City, has partnered with Georgia State University (GSU) to conduct AI marketing research aimed at enhancing restaurant operations. The research project, led by Country Fried Creative owner Joe Domaleski, MBA, and GSU data science intern Vinay Revanuru, MS, uses advanced data analysis and AI machine learning techniques to provide actionable insights for restaurant owners.
The project involved gathering online restaurant reviews, performing exploratory data analysis to identify trends, and using a Python program and a large language model (LLM) to conduct sentiment analysis. This comprehensive approach allowed the team to uncover important patterns and sentiments from customer feedback, providing a deeper understanding of consumer preferences.
Joe Domaleski, owner of CFC, emphasized the importance of this collaboration: “Partnering with GSU has been a fantastic opportunity to combine academic expertise with practical business applications. My MBA is from GSU so it seemed natural to partner with my Alma Mater on this exciting project. Anyone can use generative AI like ChatGPT to write text and generate fake images for marketing, although we wish people wouldn’t do that. We wanted to leverage AI in a deeper way that’s useful to local businesses. Our research considers customer feedback to generate insights that can help restaurants improve their marketing and operations, from menu development to customer service.”
Vinay Revanuru, the GSU data science intern who did most of the coding and implementation of the project, shared his excitement about the practical applications of their findings: “Using AI to analyze customer reviews has been incredibly rewarding. We’ve been able to identify key trends and sentiments that can guide restaurant owners in making data-driven decisions. This experience has shown me the real-world impact that data science can have on local businesses.” Originally from India, Vinay has been studying at GSU for the past two years. Vinay is finishing up his Master of Science degree in Data Science and will graduate in August. He interned with CFC this summer as part of his degree requirements.
The AI sentiment analysis project itself involved five major steps:
- Project setup – As the project sponsor, Joe had to determine the business case for the project. In the food / beverage business, reviews can make or break a restaurant. Knowing that text based reviews are traditionally hard to analyze manually, he and his marketing team at CFC decided to apply AI to that task.
- Data collection – Vinay consulted with the marketing team at CFC to select the test client, Maguire’s Irish Pub, and the online review platforms to consider – Google Business Reviews, Yelp, and TripAdvisor. He used a combination of Python, the Beautiful Soup library, Selenium, and GoLang to harvest the restaurant reviews from the web. More than 10,000 reviews per hour could be processed with the setup.
- Data analysis – The CFC team lead by Vinay did some exploratory data analysis (EDA) to summarize the quantitative aspects of the data such as the average ratings and the distribution of ratings over time – weekly, monthly, and yearly. They were able to correlate an increase in ratings to marketing campaigns that were conducted by CFC to ask customers to review Maguire’s.
- Sentiment analysis – Vinay used a Python program that tapped into the power of an AI large language model (LLM) to process the text based portion of the reviews. The goal was to get a contextual understanding of the written reviews, note any extremes, and determine the trends. Vinay expressed the results as a word cloud and word frequency chart to visually express the customer sentiments.
- Summarize the insights – once the data was analyzed quantitatively and qualitatively, the CFC team worked together to summarize the actionable insights for Maguire’s. The insights were summarized with numerical inferences about the positive and negative sentiments. Using AI, they were able to come up with recommendations for the restaurant such as, “The ambience is frequently lauded, suggesting that maintaining this strength is vital for positive customer impressions.” Translating that AI-generated statement, Joe had this to say – “What the model says is that people love the ambience at Maguire’s and so do I.”
Domaleski adds, “We chose our long-term client Maguire’s Irish Pub as the initial test case. We already knew they have great food and excellent customer service, but we wanted to see if the customer sentiments actually supported their great reputation and if there were areas of improvement. I’m pleased to report their great reviews are substantiated by the data but like any business, there’s always room for improvement. As a local business owner, I think it’s important to support other local businesses. Larger companies in our community should be aware that some really innovative research and talent is being developed right here in their backyard. Support local whenever you can. Also, support students whenever you can. Vinay Revanuru exceeded our expectations as an intern this summer. I look forward to watching his career with great interest.”
This collaborative sentiment analysis project demonstrates how data science and AI can benefit not only large businesses, but smaller local businesses like Maguire’s Irish Pub. By understanding customer sentiments and trends, business owners can tailor their offerings to better meet customer needs, enhance satisfaction, and ultimately drive business growth. For more information about this project, please contact the project sponsor – Country Fried Creative.
