SentimentLens: ML-Powered Customer Feedback Analyser
Machine Learning
Project Overview
A machine learning pipeline that ingests customer feedback from multiple channels, classifies sentiment, extracts topics, and surfaces actionable insights through a real-time dashboard — replacing manual feedback triage.
Client: Concept Build — RetailEdge Group
Duration: 6 weeks
The Challenge
The client received thousands of customer feedback submissions weekly across email, in-app surveys, and review platforms. Manual triage by a small team meant critical issues were often spotted days late. They needed automated, real-time insight surfacing.
Our Solution
Built a Python ingestion service that consumes feedback from multiple sources via Apache Kafka. A fine-tuned BERT model (Hugging Face Transformers) classifies sentiment (positive / neutral / negative) and a secondary scikit-learn topic classifier groups feedback into product areas. FastAPI exposes a REST endpoint for the React dashboard, which surfaces trending topics, sentiment shifts, and volume anomalies in real time using Recharts. The full ML pipeline and API are containerised with Docker; GitHub Actions runs model evaluation tests and deploys on every merge to main.
Results
- Feedback triage time reduced from days to near real-time
- BERT sentiment model achieved 91% accuracy on the client's domain-specific corpus
- Trending negative topics surfaced to product teams within minutes of volume spike
- CI/CD pipeline includes automated model evaluation — accuracy regressions block deployment
Client Testimonial
"Alicorn built the LiteCloud practice management platform for us — it handles our entire client workflow, task tracking and reporting. It has been running in production for over 8 months and the team has been responsive throughout. The co-founders are directly involved, which makes a real difference."
Lee Phillips
Digital Data Lead, Twinings Ovaltine
Ready to Start Your Project?
Let's discuss how we can bring your ideas to life and drive your business forward.
Get in Touch