Abstract:Knowledge graph, also known as scienti c knowledge graph, can reveal the dynamic development rules in complex knowledge elds. How to clearly present the internal structure of knowledge graph is particularly important, however, the current visualization research based on knowledge graph is rare. In this paper, varieties of data related to education are mined from massive web data, and are fused together. Then knowledge graph which is centered on educational events is constructed utilizing extracted named entities and entity relations. We construct a visual analysis platform for education knowledge graph, EduVis, which can support users to do associated analysis of education, and enable users to obtain the public opinions. In EduVis, we design and implement a) a word cloud treemap to provide an overview of education knowledge graph, b) a layout of events relation network graph based on topological structure and timeline to explore in details, c) a click tracking path to record the history of users' clicks and help users to backtrack. The case studies show that the aforementioned visual analysis methods for our knowledge graph can meet users' demands for data analysis tasks.

Abstract:We present a visual analytics system to understand the operation data of a company, GAStech, from IEEE VAST Challenge 2016. The data include proximity data recording the locations and movements of employees, and heating, ventilation, and air conditioning (HVAC) data recording the environmental conditions in the building. Analyzing the data to detect the suspicious behaviors of some disgruntled employees is of special interest. Our system provides coordinated multiple views to visualize the proximity data and the HVAC data over time. Visual hints and comparisons are designed for users to identify abnormal patterns and compare them. Furthermore, the system automatically detects and correlates the anomalies in the data. We provide use cases to demonstrate the effectiveness of our system.

Abstract:Mining frequent patterns from people’s trajectory has become a hot topic in big data research. Previously, these data mostly come from GPS. Compared with GPS data which is more densely sampled, base station data is extremely sparse in both time and space. Trajectory discovery from base station data becomes much more challenging. In this paper, we propose a new method to effectively solve this problem. In our method, we assume that the locations of objects are sampled over a long time period. First, sequential pattern mining algorithm is employed to find frequent passing areas of a person’s route every day. Second, frequent paths are pieced together by points of records which pass through frequent passing area. Finally, to ensure credibility and efficiency, we depend on the location information provided by scattered points which piece together frequent paths to mine frequent road paths.

Abstract:In this paper, we describe our progress in creating the framework for an interactive application that allows humans to actively participate in a t-SNE clustering process. t-SNE (t-Distributed Stochastic Neighbor Embedding) is a dimensionality reduction technique that maps high dimensional data sets to lower dimensions that can then be visualized for human interpretation. By prompting users to monitor outlying points during the t-SNE clustering process, we hypothesize that users may be able to make clustering faster and more accurate than purely algorithmic methods. Further research would test these hypotheses directly. We would also attempt to decrease the lag time between the various components of our application and develop an intuitive approach for humans to aid in clustering unlabeled data. Research into human assisted clustering can combine the strengths of both humans and computer programs to improve the results of data analysis.

Abstract:NBA game is one of the world’s most exciting sports. Multiple datasets for a NBA game are available, e.g., play-by-play data, shot-position data, twitter and videos. Existing methods for visualizing game data commonly focus on the composition of multi-variate information of a game process. In this paper, we introduce a new parametric modeling approach, Performance Histogram Curve (PHC), that locally and adaptively encodes the game progression with game-related features derived from the play-by-play data. By transforming a PHC into the two-dimensional space with a two-phase projection technique, we create a unique 2D line representation. The 2D representation and auxiliary views abstracts the progress of a game and the performance of each team along the timeline. Our integrated system favor browsing a single play, analyzing game performance, and comparing multiple games. We conducted two case studies to demonstrate the effectiveness of our approach.

Abstract:We present a novel tool to visualize dependency trees in a hyperbolic layout, and to provide visual support for comparative evaluation of parsing errors. Compared with traditional flat tree visualization, our hyperbolic tree visualization tool can be more convenient for showing long-range dependencies. Our tool integrates the hyperbolic view with a flat view, and support corpus-level error analysis. It offers several features, including statistical analysis of error distributions, visual analysis of individual dependency trees, and an integrated online interface.