🛍️ NLP Capstone Project: Customer Review Analysis – Women’s Clothing E-Commerce 📌 Project Overview This project analyzes over 23,000 customer reviews for a leading women’s clothing e-commerce platform. The goal is to extract meaningful insights using Natural Language Processing (NLP), understand customer sentiment, and build predictive models for recommendation and rating prediction.

🎯 Business Objectives 🔍 Perform exploratory data analysis (EDA) on customer reviews and demographics.

💬 Conduct text mining to identify frequent words in positive and negative reviews.

🌍 Analyze sentiment trends by category, location, age group, and channel (Web/Mobile).

🤖 Build predictive models to:

Classify customers likely to recommend a product.

Predict product ratings based on review text.

🧠 Perform topic modeling to uncover themes in customer reviews.

📁 Dataset Details Column Name Description Product ID Unique ID of the product Category Product category Subcategory1 Subcategory level 1 Subcategory2 Subcategory level 2 Location Customer’s geographic location Customer Age Age of the customer Channel Purchase channel (Web or Mobile) Review Title Short title for the review Review Text Full review text Rating Rating given by the customer Recommend Flag Whether the customer recommends the product (Yes/No)

🧪 Techniques & Tools Used Programming Language: Python

Libraries: Pandas, NumPy, Matplotlib, Seaborn, Plotly

NLP: NLTK, spaCy, WordCloud, scikit-learn, TextBlob

Modeling: Logistic Regression, XGBoost, Random Forest

Topic Modeling: LDA (Latent Dirichlet Allocation)

Dashboarding: Streamlit / Plotly Dash (optional)

📊 Key Features 📈 EDA Dashboard: Category-wise review count, age distribution, recommendation trends

🌐 Word Clouds: For positive vs negative reviews

📊 Sentiment Analysis: By channel, location, product category, and age group

🔍 Topic Modeling: Extracting key themes from customer feedback

🧠 Classification Models: Predicting recommendation flag

🎯 Regression Models: Predicting customer rating from review text

🚀 How to Run Clone the repo

bash Copy Edit git clone https://github.com/rahulraimau/womens-clothing-nlp.git cd womens-clothing-nlp Install dependencies

bash Copy Edit pip install -r requirements.txt Run the Jupyter notebook or Streamlit app

bash Copy Edit jupyter notebook

or

streamlit run app.py 📌 Sample Visualizations WordCloud of Top Positive/Negative Words

Sentiment by Age Group

Top Recommended Categories

Topic Clusters per Product Type

📦 Outputs sentiment_model.joblib: Trained sentiment classification model

rating_predictor.pkl: Regression model for rating prediction

lda_model.pkl: Topic model with top keywords

final_dashboard.html: Interactive dashboard output (optional)

👤 Author Rahul Rai 📧 rahulraimau5@gmail.com 🔗 GitHub | Kaggle | HackerRank