Data-driven construction of visual query interface for time series data

Abstract

This Final Year Project explores data-driven techniques to help end-users to query such time series data. The goal of the project is to design and implement an efficient and scalable algorithm that can fill up pattern panels in SENSOR, a novel data-driven sketch-based visual query interface (VQI) construction system that automatically constructs its VQI from time series data using representative objects called shapelets, which are formed from the data. With a multi-optimization objective of High Diversity, High Coverage and Low Cognitive Load, methods to interpret and work with time series data and various models have and are being explored to facilitate the development of the final algorithm.

This Final Report aims to walk through the following components of the project – an introduction, literature review of existing work and new research, an analysis of its requirements with review. Following this, this report also covers the proposed algorithm in detail, a walkthrough on how the output will be visualized, usability tests, and finally future work to further improve upon the current work.