We introduce a causal knowledge graph (KG) of drug-condition relationships extracted from patient-authored reviews leveraging generative AI. While the graph construction is based on the JSL-MedLlama large language model opportunely fine-tuned on the MIMICause dataset, the core contribution of this ongoing work is the design and deployment of an interactive dashboard that enables users to visually explore, analyze, and query the KG content. The dashboard features faceted navigation, frequency-based entity analytics, and chord diagrams that highlight the strength of causal relations across five categories: Cause, Prevent, Enable, Hinder, and Other. The KG contains over 3,000 distinct triples, with drug and condition entities linked to biomedical ontologies, ensuring semantic alignment and linked open data compatibility. The KG is published in RDF format, and accessible via a SPARQL endpoint, along with a developed user-friendly dashboard and ad-hoc visualizations. By offering rich visual and programmatic access to patient-derived causal knowledge, our system supports research in pharmacovigilance, biomedical NLP, and knowledge-driven healthcare applications.