Create Your First Project
Start adding your projects to your portfolio. Click on "Manage Projects" to get started
Customer Review Sentiment Analysis and Product Recommendation Insights
Date
April 2023
Coding Assignment
This project uses Python and various data science techniques to analyze customer reviews for understanding sentiment, reviewer patterns, and product preferences. Key tools include NetworkX for visualizing product connections, NLTK for sentiment analysis, and KMeans for clustering sentiment patterns.
Google colab
Problem Statement
Objective: “To analyze customer reviews for sentiment, product rating trends, and reviewer behavior to generate insights into customer satisfaction and recommend popular products. The analysis used graph visualization and sentiment analysis to identify review patterns.”
Data Collection and Methodology
Data Source: Customer reviews were sourced from a JSON dataset, containing details such as product IDs, ratings, review text, and reviewer IDs.
Data Processing:
Parsed JSON data to extract review information.
Processed text data for sentiment analysis using NLTK’s Vader sentiment analyzer.
Used KMeans clustering to identify patterns in sentiment and review count.
Key Analyses and Visualizations
Product Review Network:
Visualization: A network graph displaying connections between products frequently reviewed by the same customers.
Insight: Reveals relationships among popular products, offering insights into co-purchase behaviors.
Sentiment Analysis:
Method: Used Vader to determine positive, negative, or neutral sentiment in each review.
Result: Quantified sentiment scores to assess overall customer satisfaction, identifying products with the highest and lowest sentiment.
Clustering by Sentiment and Review Count:
Method: KMeans clustering grouped products based on average sentiment and review volume.
Insight: This analysis highlights popular products and those with positive or negative feedback trends, helping to identify product strengths and areas for improvement.
Technical Challenges and Solutions
Handling Large Text Data: Preprocessed review text for efficient sentiment analysis.
Visualizing Relationships: Employed NetworkX for constructing and simplifying complex product-reviewer graphs.
Clustering Tuning: Adjusted KMeans parameters for meaningful clustering of sentiment and volume.
Tools and Technologies Used
Languages & Libraries: Python, NetworkX (for graph analysis), NLTK (for sentiment analysis), Matplotlib (for plotting), KMeans (for clustering).
Platform: Google Colab for data processing and visualization.
Learning Outcomes
Gained experience in sentiment analysis and clustering techniques.
Developed proficiency in graph visualization to illustrate relationships in customer behavior data.
Enhanced my understanding of data preprocessing for large text-based datasets.
