NLP for regulatory text

Security Policy & Compliance Automation

A group project that reads privacy-policy documents and scores them against GDPR requirements, with section-by-section analysis and gap identification.

The problem

Checking a policy document against GDPR, HIPAA or PCI DSS by hand is slow and easy to get wrong. The project automates extraction and classification of the relevant policy statements.

Approach

• Sentence-level processing with NLTK and spaCy, including named entity recognition.
• SentenceTransformers embeddings with PyTorch classification models for GDPR compliance.
• A Streamlit interface for uploading documents and reviewing results, with Plotly dashboards.
• Exportable reports with compliance scores and identified policy gaps.

Stack

• SentenceTransformers
• spaCy
• PyTorch
• Streamlit
• Plotly