Skills: Python libraries (Matplotlib, Numpy, Pandas, Seaborn), Gemini AI, Google Colab

Key Objectives

• Collect and clean a dataset of 1K+ Amazon reviews for various headphones, including that of Beats

• Analyze the dataset using statistics and visualizations

• Conduct sentiment analysis on the dataset through traditional sentiment analysis and AI

• Summarize analyses and findings into an actionable plan for Beats

Project Description

For this externship capstone project, I utilized consumer insight analysis to provide actionable product recommendations for Beats by Dre. First, I compiled and cleaned a dataset of 1000+ Amazon reviews spanning 10 various products to identify market opportunities for Beats. This was done using Python libraries such as Pandas and Numpy, as well as Oxylabs since the Extern.com-provided Amazon scraper did not work. Afterward, I conducted exploratory data analysis (EDA) using the same libraries and Matplotlib to get a sense of how Beats headphones were performing compared to its competitors.

To gain deeper insights, I conducted sentiment analysis using two different methods: Textblob and Gemini AI. Using TextBlob, I classified product reviews by 3 different sentiment categories (positive, neutral, negative) and aggregated the results for Beats and their competitors. I created Gemini AI prompts to holistically analyze the entire dataset, asking the AI to pose as product manager and identify Beats’ problems and weaknesses, enjoyable traits, and top customer recommendations for improvement.

At the end of my externship, I documented my findings and processes into a comprehensive 10-part Google Colab report with references, utilizing both code and text cells to display my work. If I redo this project, I could probably conduct a deeper EDA and sentiment analysis, as well as randomly sample my product reviews to provide a more accurate depiction of Beats’ performance.

To protect against plagiarism from future program participants, I didn’t display the display project, but here are some screenshots in no particular order.