CustomerInsight

AI-powered customer review analysis platform

Live Demo →

---

Overview

CustomerInsight is an interactive analytics platform that uses NLP and machine learning to extract actionable insights from customer reviews. Built with Streamlit, it provides real-time sentiment analysis, keyword extraction, topic modeling, and anomaly detection — optimized for both Chinese and English text.

Features

• Sentiment Analysis — BERT-based sentiment classification with confidence scores. Uses roberta-base-finetuned-jd-binary-chinese for Chinese and bert-base-multilingual-uncased-sentiment for English.
• Keyword Extraction — TF-IDF keyword extraction with word cloud visualization, trend analysis, and rating-based comparison.
• Topic Modeling — LDA and K-Means clustering with topic network graphs, heatmaps, and trend tracking.
• Anomaly Detection — Isolation Forest-based outlier detection with multi-feature analysis (rating, text length, sentiment).
• Interactive Filtering — Filter by date range, rating, text length, and keywords with real-time updates.
• Data Export — Download filtered data and analysis results as CSV.

Quick Start

# Clone the repository
git clone https://github.com/ChanMeng666/customer-insight.git
cd customer-insight

# Install dependencies
pip install -r requirements.txt

# Run the application
streamlit run app.py

Then open http://localhost:8501 and upload the sample dataset from data/example_dataset.csv.

Data Format

Column	Type	Required	Description
timestamp	datetime	Yes	Review timestamp
content	string	Yes	Review text
rating	float	No	Rating (1-5 scale)
category	string	No	Review category
user_id	string	No	User identifier

Architecture

graph TD
    A[CSV/Excel Upload] --> B[Data Processor]
    B --> C[Text Cleaning]
    C --> D{Analysis Engine}
    D --> E[Sentiment Analyzer]
    D --> F[Keyword Analyzer]
    D --> G[Topic Analyzer]
    D --> H[Insight Analyzer]
    E --> I[Visualization Layer]
    F --> I
    G --> I
    H --> I
    I --> J[Streamlit Dashboard]

Loading

Tech Stack

Framework: Streamlit · Language: Python 3.9+

NLP & ML: Transformers · PyTorch · jieba · NLTK · scikit-learn · Sentence Transformers

Visualization: Plotly · Matplotlib · WordCloud · NetworkX

Data: Pandas · NumPy · SciPy

Contributing

See CONTRIBUTING.md for the development guide.

License

MIT © Chan Meng