This report offers insights from a comprehensive analysis of restaurant reviews, aiming to enhance customer satisfaction and business performance. Leveraging advanced Natural Language Processing (NLP) techniques, we have processed and analyzed a dataset of customer reviews to discern patterns related to customer sentiment. The purpose is to identify factors influencing positive and negative reviews, thereby enabling targeted improvements in service quality and customer experience.
Our primary objectives can be compartmentalize in three: 1. To decode customer sentiment and preferences: Understanding what delights and dissatisfies customers across different restaurant experiences. 2. To identify operational and service excellence areas: Pinpointing service aspects that consistently correlate with positive reviews. 3. To tailor marketing and operational strategies: Leveraging insights to refine marketing approaches and operational efficiencies.
The analysis employed several machine learning models to classify reviews into positive and negative sentiments. Key findings include:
-
Model Performance: The SGDClassifier model outperformed others, achieving a 90% accuracy rate in predicting customer sentiment. This high level of accuracy demonstrates the model's effectiveness in distinguishing between positive and negative reviews based on textual content.
-
Sentiment Insights: Word cloud visualizations highlighted prevalent themes in positive and negative reviews, offering direct feedback on what customers appreciate or dislike. Positive reviews frequently mentioned attributes such as "friendly staff," "excellent service," and "delicious food," while negative reviews often cited "long wait times," "poor service," and "average food."
-
Actionable Strategies: Based on the model's predictions and the sentiment analysis, businesses can focus on enhancing staff training, improving food quality, and optimizing service efficiency. Additionally, recognizing specific issues mentioned in negative reviews allows for targeted interventions to address customer complaints effectively.
| Metric | SGDClassifier | XGBClassifier |
|---|---|---|
| Accuracy | 90.37% | 89.77% |
| Precision | 90.21% | 89.53% |
- Saad Umar
- Yiwen Mei
- Adil Ashraf
- Jake Gu