Towards Semantic Search in Building Metadata

This report presents the design of a search engine that can be used to search across internet-of-things sensor data and metadata in the context of building management. Search engines are most commonly used to allow humans to retrieve relevant web-pages from the internet given a natural language query. The system presented here also takes natural language queries as input but instead of returning web-pages, its primary function is to return visualizations of time-series data in the form of line charts. The system also allows users the ability to name visually-defined phenomena in the returned line charts and search across sensor data for similar data shapes.