Customer Sentiment Dashboard

A full-stack, professional dashboard for qualitative customer feedback analysis. This application allows teams to process bulk unstructured text reviews and transform them into actionable intelligence, featuring timeline mapping, sentiment distribution, and semantic keyword extraction.

• Batch Processing: Paste large volumes of unstructured qualitative reviews for instant processing and semantic aggregation.
• Sentiment Trend Mapping: Visualizes customer satisfaction over inferred timelines using interactive moving average charts.
• Keyword Density Analysis: Automatically extracts and categorizes high-frequency praise and complaint terminology into visual word clouds.
• Automated Executive Summaries: Generates natural language summaries and identifies the top 3 actionable areas for improvement based on the dataset.
• Customizable Thresholds: Dynamically adjust the scoring thresholds for "Positive" and "Critical" sentiment to tailor the health check metrics to your specific organizational standards.
• Data Export: Built-in tools to export intelligence reports to PDF for stakeholder distribution, or download raw structured data to CSV.

• Frontend: React 19, TypeScript, Tailwind CSS
• Data Visualization: Recharts, D3.js (d3-cloud)
• Backend: Node.js, Express
• Analysis Engine: Google Gen AI SDK integration for semantic processing and structural insight extraction
• Tooling: Vite, ESBuild

• Node.js (v18 or higher recommended)
• npm or yarn

1. Clone the repository and install dependencies:

   npm install

2. Configure environment variables. Copy the example .env file and add your API keys:

   cp .env.example .env

   Note: Ensure you populate GEMINI_API_KEY for the semantic analysis engine to function properly.

Development Mode: Starts the Vite development server with the Express backend integration.

npm run dev

Production Build: Compiles the React frontend and bundles the Express backend into a single target.

npm run build

Start Production Server: Runs the optimized compiled output.

npm start

1. Launch the Application: Navigate to the dashboard UI.
2. Input Data: Paste a set of unstructured text reviews into the analysis text area, or use the "Load sample reviews" option for a quick demonstration.
3. Generate Intelligence: Click to process the batch. The backend securely parses and categorizes the sentiment.
4. Review Metrics:
   • Analyze the "Health Check" breakdown.
   • Use the threshold sliders to fine-tune categorization parameters.
   • Review the Executive Summary and Top Actionable Areas.
5. Export & Share: Use the "Export PDF" or "Export CSV" tools in the top right to save and share your findings.

This project is licensed under the Apache 2.0 License.