Data Visualization in Educational Datasets using a Rule-Based Inference System

Dynamic data visualization can be a very useful analytical tool for discovering insights in complex real-world data sets with high dimensionality and large variety of data types. We leverage publicly available data sets from Washington State's Public Education System to demonstrate usefulness of data visualization as a tool for analysis and effective decision making. We created "querybuilder", a web-based interface which allows the user to generate ad-hoc queries. Our online inference system efficiently generates dynamic visualizations for user-specified queries. We then address the question of how to select an appropriate visualization type that would be best suited for the result of a specific query on the given data set. The main motivation for this work is developing a rule based inference system to automatically select the appropriate visualization type.