Semantic Querying for Data Science on the Edge
Project Overview
The rapid proliferation of Internet of Things devices in smart homes generates vast quantities of personal data, much of which is transmitted to cloud platforms for processing. This raises significant privacy concerns and introduces unnecessary network overhead. This project envisions future smart homes acting as personal data stores that serve privacy-aware analytics entirely at the edge. The prototype repurposes several semantic web ontologies, including the Semantic Sensor Network ontology, to model IoT products and services running on the openHAB home automation platform. By executing data wrangling operations at the edge rather than in the cloud, the system minimises data acquisition and empowers data scientists to request only the specific data streams they need, thereby reducing unnecessary data transfers. The approach leverages OWL ontologies to provide a structured knowledge representation of available sensors, their capabilities, and the data they produce.
The semantic layer enables intelligent querying and filtering of IoT data before it leaves the home network. Data scientists can discover what data is available and request only the streams relevant to their analysis, ensuring that unnecessary personal information is never transmitted beyond the home.
The work demonstrates that edge-based semantic technologies can effectively support personal data science workflows. By keeping data processing local, the system preserves user privacy and reduces bandwidth consumption in smart home environments. This approach offers a practical pathway toward privacy-aware smart home analytics without sacrificing the richness of data science capabilities.
Team
