SentimentRadar | Devpost

Inspiration

SentimentRadar was inspired by the need for telecom companies to move from reactive to proactive customer service. By combining sentiment analysis, AI diagnostics, and geographic visualization, it empowers teams to detect and address satisfaction issues in real time before they escalate.

What it does

SentimentRadar visualizes customer sentiment on an interactive US heatmap, tracks happiness scores, and correlates them with network outages. It uses AI-driven diagnostics to identify root causes, recommend internal resolutions, and even generate customer-facing updates in real time.

How we built it

The app uses Next.js 14, React, and TailwindCSS for the frontend, with Leaflet.js for map visualization and Supabase for data storage and APIs. AI components like NVIDIA Nemotron and HuggingFace models power multi-step sentiment and diagnostic workflows, creating a scalable, real-time analytics system.

Challenges we ran into

We faced challenges with heatmap calibration, client-side rendering in Next.js, and data aggregation across timestamps. Building smooth real-time updates and designing a modular AI workflow also required careful optimization and system design.

Accomplishments that we're proud of

We successfully integrated large-scale sentiment analysis with diverse regional datasets to generate a clear, intuitive visualization of customer satisfaction across U.S. cities. This fusion of geographic and emotional data allows users to instantly pinpoint problem areas and understand trends at a glance.

What we learned

We learned how to merge sentiment, geographic, and temporal data into a cohesive real-time visualization that’s both insightful and actionable. Our team deepened its understanding of structuring AI-driven workflows, optimizing data pipelines, and designing responsive dashboards that handle live updates efficiently. This project also taught us the value of balancing clarity, performance, and interpretability when working with large-scale analytics systems.

What's next for SentimentRadar

We plan to enable automated workflows that can issue support tickets to external platforms and trigger small, intelligent responses when customer sentiment drops. This will transform SentimentRadar from a monitoring dashboard into an active system for real-time incident response and customer experience management.