PharmaSignal

Inspiration

Pharma companies rely heavily on manual monitoring to understand how medicines are perceived after release. We noticed that early signs of safety concerns or trust issues often appear as small, scattered discussions online but are detected too late. This inspired us to build a system that can surface these weak signals early and support proactive decision-making.

What it does

PharmaSignal is a web-based AI decision-support platform that analyzes public discussions from multiple sources to detect early warning signals related to drug safety, trust, and adoption risks. It provides confidence-scored insights and trend visualizations for internal pharma teams, without making automated or medical decisions.

How we built it

We built a React.js frontend for visualization and interaction, a Node.js backend for API handling, and Python-based services for NLP and data analysis. Lightweight machine learning models filter relevant data, detect emerging patterns, and generate explainable insights that are presented through a dashboard.

Challenges we ran into

The biggest challenge was handling noisy and unstructured public data while avoiding false signals. Another challenge was designing the system to assist human decision-makers without over-automating sensitive decisions.

What we learned

We learned how to design responsible AI systems, handle noisy real-world data, and build scalable web architectures that balance automation with human oversight.

What's next for PharmaSignal

Next, we plan to add more data sources, improve multilingual support, enhance signal confidence scoring, and integrate the platform with existing pharma monitoring workflows.