课程摘要 : 数据可视化主要旨在借助于人的视觉感知通道，通过把数据转化为图形图像，清晰有效地传达与沟通信息，是数据探索分析的重要手段。我们将综述可视化的历史沿革和重要的基本原理。
课程摘要 : 本节课介绍可视化面向不同类型数据所发展的不同构型设计，包括表格类型、高维数据、网络数据、时空数据等不同类型对应的构型。
课程摘要 : This talk explores a concept-driven process and method for creating effective and interesting visualizations. Coupled with a few hand-on exercises, we’ll cover generating suitable, exciting concepts and ideas for visualization; collecting and examining fitting data; developing and selecting appropriate, effective presentation; designing and enhancing the aesthetics and impact of final outcome. We will also broadly discuss how to use data visualization as a means for communicating ideas and stories through review of various methods and approaches for visualizing data. This talk aims to help both artists/designers and data researchers to expand their approach to visualization, hopefully to achieve certain balance among aesthetic quality, design-thinking and statistical rigor in their work.
课程摘要 : 数据可视化中，数据集通常很大，一般情况下不存在能包括全部数据信息的视图。交互提供了用户与计算机交流的渠道，面对不同的分析任务，用户可以调整相应的可视化结果。如何有效地设计交互，是数据可视化中非常重要的部分。
课程摘要 : 基于北大可视化实验室的优秀项目，本次课程从案例分析的角度介绍优秀可视化的各个要素。
课程摘要 : 本次课程研讨初学者如何开展可视化研究。
课程摘要 : 本节课程讲授国内外可视化的前沿进展和未来趋势。
课程摘要 : Movement is a key element of human activities. The understanding of movement can benefit many applications including traffic management and urban development. Hence, movement visualization has always been a hot research topic in cartography, GIS, and transportation. However, as increasing amounts of movement tracking data are being collected, conventional visualization methods such as space-time cube and flow map, can easily cause significant visual clutter and occlusion issues. This talk reports recent advancements in visualization, selection, and analysis methods for movement data, followed by an in-depth discussion of their applicability and limitations. Assorted examples on urban traffic data will be presented to demonstrate real-world usage scenarios.
课程摘要 : Interacting with people to obtain first-hand insights is a critical practice that occurs throughout a human-centered design process. Across four lectures with in-class exercises, we will introduce the key concepts, common tactics, and important lessons in the design and execution of various types of user studies, such as needfinding, iterative design feedback, and systematic evaluation. We will also review some basic methods to analyze and interpret user study results.
课程摘要 : The Intel Rendering Framework is a suite of open-source libraries and frameworks for large-scale high-fidelity visualization on Intel architecture. It includes the Embree ray tracing kernels, the OSPRay ray tracing API for visualization, the OpenImageDenoise library and OpenVKL volume kernel library. In this talk we will cover recent advancements including distributed frame buffer functionality, path tracing materials, unstructured mesh volumes, Tapestry for web-based streaming visualization, OSPRay Studio for lightweight data exploration, and full integration in Kitware’s ParaView platform.
课程摘要 : This event will consist of a hands-on walkthrough of several demos using the Intel OSPRay ray tracing API for scientific visualization, a tutorial for end-users on how to easily use OSPRay in Kitware’s Paraview platform, and a break-out session.
课程摘要 : With the increasing availability of computational data analysis and modelling tools that can be utilised out-of-the-box, the route from data to results is now much shorter. However, these advancements also come together their own limitations, and a data scientists need to be aware of the pitfalls and act carefully to question every observation and method used within each step of the data analysis process. Visual analytics approaches where interactive visualisations are coupled tightly with the algorithms offer effective methodologies in conducting data science in such inquisitive, rigorous ways. This tutorial will discuss how visual analytics can facilitate such practices and will look at examples of research on how data can be transformed and visualised creatively in multiple perspectives, on how comparisons can be made within different models, parameters, and within local and global solutions, and on how interaction is an enabler for such processes. We will investigate how interactive data analysis techniques can transform the data analysis pipeline from the wrangling to the modelling stage. We will discuss a series of techniques that are specifically developed for working with high-dimensional and spatio-temporal datasets. We will try to demonstrate how visual analytics fosters a more comprehensive, richer modelling process by facilitating an inquisitive, reflective and critical approach to doing data science and how it enables us to ask the right questions.
课程摘要 : 随着当今计算性能的高速发展，科学模拟计算在各科学研究、工程实践领域中的应用越来越广泛，而其起到的作用亦与日俱增。然而，由于数据本身的不可见性，研究人员必须借助可视化工具观察、理解数据中蕴含的物理现象，从而达到验证科学假设乃至提出新的科学构想的目的。当前科学模拟计算呈现规模大、复杂度高等特征，给科学家通过视觉系统感知可视化传递的信息造成了极大困难，也为可视化研究带来了重大挑战。在此报告中，我将讨论如何通过在科学可视化中引入交互式探索手段，从而帮助科学家在观察数据的过程中将展示的数据化繁为简，从大规模复杂数据中发掘关键特征及其内在联系。我将结合流场数据以及时变多变量数据这两类常见的科学数据，介绍这些方法及理解科学数据中的应用。