Description: The first phase of this project was building the PKM (Personal Knowledge Management) foundation in Notion and Obsidian for the WASH sector, specifically waste management. In Notion, every document I read is summarized and stored in a structured database, complete with metadata like source, geography/region, date, topic tags, and key findings, making it easy to filter and sort as needed. Meanwhile, in Obsidian, the same notes are kept in markdown format and visualized through a graph view, so the connections between topics or issues become much easier to see. This phase became the foundation before eventually evolving into a RAG-based chatbot in the next phase.
Description: The second phase of this project was developing the PKM built in phase one into a RAG-based (Retrieval-Augmented Generation) chatbot, deployed via Streamlit so it can be accessed directly from a browser. This chatbot uses 119 summarized documents on waste management as its knowledge base, and offers four modes: Chat (RAG) for direct Q&A, with answers automatically matching the language of the question, Semantic Search for finding documents by meaning, Count & Filter for counting and filtering documents by metadata, and Latest/Oldest for browsing the newest or oldest entries. Through this phase, notes that used to just sit there, saved and manually tagged, can now actually be discussed directly.