Multi-Attribute Similarity Search

SimSearch is a tool that simplifies data exploration by enabling top-k similarity search over large collections of entities involving multiple heterogeneous attributes from different sources. It supports different modes for data access and query over diverse types of attributes, including textual, numerical and spatial. Users can specify their query parameters and preferences through a web interface, and visually inspect and compare the results through appropriate visualizations for the different types of attributes involved.

The code of SimSearch is publicly available on GitHub, including both the backend and the UI.

More information about SimSearch can be found in the demo paper “Multi-Attribute Similarity Search for Interactive Data Exploration” by Kostas Patroumpas, Alexandros Zeakis, Dimitrios Skoutas and Roberto Santoro, which was presented at the SIMPLIFY Workshop colocated with the EDBT Conference on March 23, 2021, as well as in the video below.