The primary objective is to gather, clean, and analyze public-domain data on traditional medicinal herbs and formulations to prepare a structured training corpus for AHI. Students will identify credible online sources, extract herb names, formulations, and usage guidelines, normalize and merge data into a unified schema, and conduct exploratory data analysis (EDA) to reveal key patterns—such as top herbs by frequency, co-occurrence of ingredients and ailments, and geographic usage trends. The final output will support AHI’s next fine-tuning cycle by supplying high-quality CSV/JSON datasets and visual insights

Students will deliver a set of cleaned datasets (CSV and JSON) containing standardized tables for herbs, formulations, and usage cases complete with source metadata, accompanied by a brief README that outlines field definitions and normalization decisions. They will also produce a concise exploratory data analysis report (4–6 pages) featuring key visualizations (e.g., top herbs bar chart, herb-ailment heatmap, regional trends) and actionable recommendations. All data-cleaning and analysis steps will be documented in commented Python notebooks or Excel workbooks, and students will provide exported chart files (PNG/JPEG) alongside a one-page reflection note summarizing challenges, insights, and how the dataset supports AHI’s training pipeline.