https://github.com/amirhosseinhonardoust/customer-sentiment-intelligence-platform

An enterprise-grade NLP + Streamlit + SQL platform for analyzing customer feedback. Performs automated sentiment detection, stores labeled reviews in SQLite, and delivers real-time dashboards with probability insights to support business, marketing, and product optimization decisions.
https://github.com/amirhosseinhonardoust/customer-sentiment-intelligence-platform

community-project cost-of-living dashboard data-analysis data-visualization economic-analysis inflation-tracking local-data open-data pandas price-tracker public-insight python sqlite streamlit

Last synced: 2 months ago
JSON representation

An enterprise-grade NLP + Streamlit + SQL platform for analyzing customer feedback. Performs automated sentiment detection, stores labeled reviews in SQLite, and delivers real-time dashboards with probability insights to support business, marketing, and product optimization decisions.

• Host: GitHub
• URL: https://github.com/amirhosseinhonardoust/customer-sentiment-intelligence-platform
• Owner: AmirhosseinHonardoust
• License: mit
• Created: 2025-10-29T09:30:31.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2025-10-29T09:51:08.000Z (8 months ago)
• Last Synced: 2025-10-29T11:37:50.805Z (8 months ago)
• Topics: community-project, cost-of-living, dashboard, data-analysis, data-visualization, economic-analysis, inflation-tracking, local-data, open-data, pandas, price-tracker, public-insight, python, sqlite, streamlit
• Language: Python
• Homepage:
• Size: 32.2 KB
• Stars: 1
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Customer Review Sentiment Intelligence Platform (NLP + Streamlit + SQL)

The **Customer Review Sentiment Intelligence Platform** is a production-ready analytics application that combines **Natural Language Processing (NLP)**, **SQL**, and **interactive dashboards** to deliver actionable insights on customer feedback trends.  

Designed for content, marketing, and operations teams, it provides **data-driven sentiment intelligence** to enhance decision-making, brand perception tracking, and product quality analysis.

---

## Executive Summary

Organizations generate thousands of customer reviews daily, yet most remain underutilized.  

This solution bridges that gap, automatically classifying reviews as *positive* or *negative*, quantifying sentiment confidence, and visualizing feedback patterns over time.

**Business Impact:**

- Optimize product and service strategies based on real-time sentiment signals.

- Identify pain points and satisfaction drivers per product category.

- Streamline performance reporting with ready-to-present dashboards.

- Enable faster decision-making through AI-powered review analysis.

---

## System Architecture

```text

                 +------------------+

                 |  Raw Review CSV  |

                 +------------------+

                          |

                          v

                 +------------------+

                 |  Data Ingestion  |

                 | (ETL via SQLite) |

                 +------------------+

                          |

                          v

                 +------------------+

                 |  NLP Processing  |

                 |  (TF-IDF + LR)   |

                 +------------------+

                          |

                          v

                +---------------------+

                |    Model Storage    |

                | (Joblib Artifacts)  |

                +---------------------+

                          |

                          v

                +---------------------+

                |    Streamlit UI     |

                | Real-time Analytics |

                +---------------------+

```

---

## Repository Structure

```

customer-sentiment-intelligence/

├── app/

│   └── streamlit_app.py

├── data/

│   └── reviews.db

├── models/

│   ├── model.joblib

│   └── vectorizer.joblib

├── src/

│   ├── etl_loader.py

│   ├── preprocess.py

│   └── train_model.py

└── requirements.txt

```

---

## Core Capabilities

- **Automated Sentiment Detection** | Real-time text classification using TF-IDF + Logistic Regression.  

- **Interactive Review Exploration** | Filter and visualize feedback by product, time, or rating.  

- **Confidence-Based Scoring** | Probability-weighted results for transparent interpretation.  

- **Integrated SQL Backend** | All processed reviews are persisted in SQLite for auditability.  

- **Scalable Architecture** | Modular design ready for deployment to cloud or Docker environments.  

---

## Technical Overview

| Layer | Description |

|-------|--------------|

| **Data Source** | CSV or API-based customer reviews |

| **ETL Process** | Data normalization, cleaning, and SQL ingestion |

| **Feature Engineering** | TF-IDF vectorization |

| **Modeling** | Logistic Regression (binary sentiment) |

| **Visualization** | Streamlit UI + Plotly charts |

| **Persistence** | SQLite database with labeled review storage |

---

## Visual Overview

### User Interface  



---

### Review Analytics Explorer  



---

### Sentiment Probability Distribution  



---

### Review Results Table  



---

## Deployment Guide

### Local Setup

```bash

git clone https://github.com/yourusername/customer-sentiment-intelligence.git

cd customer-sentiment-intelligence

python -m venv venv

venv\Scripts\activate  # (Windows)

source venv/bin/activate  # (macOS/Linux)

pip install -r requirements.txt

streamlit run app/streamlit_app.py

```

### Cloud Deployment (Optional)

- Package with **Docker** and deploy via **Streamlit Cloud**, **Render**, or **Azure Web Apps**.  

- For enterprise environments, integrate SQLite → PostgreSQL → Power BI pipeline for advanced analytics.

---

## Data Flow Summary

1. **Ingest Data:** Upload or connect to raw review sources (CSV or API).  

2. **Clean Text:** Tokenization, stopword removal, lemmatization.  

3. **Model Application:** TF-IDF transforms text; logistic regression predicts sentiment.  

4. **SQL Storage:** Save predictions for traceability.  

5. **Visualization:** Streamlit renders metrics, histograms, and review tables.

---

## Example Insights

- 67% of reviews show **positive sentiment**, clustered at **0.8+ probability**.  

- 33% are **negative**, primarily related to logistics and product usability.  

- High-confidence classifications indicate strong model performance.  

- Balanced feedback supports credible brand engagement insights.

---

## Governance & Compliance

- Follows **PEP8** coding standards.  

- Model artifacts tracked via reproducible pipelines.  

- SQLite ensures full audit trail for all predictions.  

- Easily extendable to comply with **GDPR** or internal data retention policies.

---

## Future Roadmap

- Introduce **Neutral sentiment** classification.  

- Add **Aspect-level sentiment** (e.g., “delivery speed”, “customer service”).  

- Enable **real-time feedback API** integration for live review analysis.  

- Extend with **topic clustering** and **keyword extraction**.