Projects

Fast Fashion Sentiment Analysis

Visit Website

A full end-to-end study that contrasts fast-fashion chains with other fashion retailers across store presence, customer sentiment, product categories, and operating practices. Using real-world data and NLP techniques, the project surfaces insights to guide brands, analysts, and researchers in understanding U.S. fashion-market dynamics and improvement areas.

Category

May 15, 2024

NLP & Machine Learning

Services

May 15, 2024

Text Mining, Sentiment Analysis, NLTK/Scikit-learn

Client

May 15, 2024

Arizona State University

Year

May 15, 2024

2025

📊 Project Overview

Datasets analysed: U.S. store locations & Google reviews
Methods & tooling: Python (Pandas, NumPy, Matplotlib, Seaborn) + NLP (NLTK, BERT via 🤗 Transformers)
Analytic focus areas
1. Store distribution across states
2. Review-based sentiment scoring
3. Category breakdown (Women’s, Men’s, Kids’, Accessories)
4. Average ratings & review volumes
5. Weekly operating-hour patterns
6. Topic modelling & word-frequency analysis

🔍 Key Insights

Theme	Key Takeaways
Store strategy	Fast-fashion brands outnumber competitors but trail in customer satisfaction.
State opportunities	📈 Growth potential identified in PA, NJ, and AZ.
Customer sentiment	Fast-fashion avg rating 2.79⭐ vs. other brands 3.64⭐. Reviews focus on price & variety; negatives cite quality and service.
NLP findings	Fast-fashion reviews cluster around “price”, “cheap”, “return”; traditional retail around “tailoring”, “bespoke”, “sustainability”.

🧠 Tech Stack

Layer	Tools & Libraries
Language	Python (Jupyter Notebook)
Data wrangling	Pandas, NumPy
Viz	Matplotlib, Seaborn
NLP / ML	NLTK, Scikit-learn, BERT
Outputs	Custom plots (bars, KDEs), topic-model dashboards