cool hit counter Dry share: how to learn relational network visualization?_Intefrankly

Dry share: how to learn relational network visualization?


Source: Yu Zhengyan

Relational network visualization of node-link: network science of points and lines)

Remove my third semester (Fall 2017) of enrollment in the Master of Fine Arts in Information Design and Visualization program at Northeastern University taking four courses, one of which focused on Relationship Network Visualization (exploration and practice of network visualization, also known as graph visualization) is part of a PhD course in the Department of Computer Science, taken twice a week.

Cody Dunne

The instructor isCody Dunne, M.S. and Ph.D. degrees from the University of Maryland, learn fromBen Shneiderman great god。Cody He focuses on information visualization、 network science、 Human-computer interaction, etc., Before joining the faculty at Northeastern, He has served asIBM Scientists in research units。 This course is accompanied by two teaching assistants( The areas of study are Relationship Network Visualization), Responsible for resolving assignment issues and other course needs。

This class is the most stressful one I have taken in the year and a half I have been in the US, because it is part of a PhD course in the Computer Science Department, so there is a lot of literature to read; in addition, some of the assignments are very demanding in terms of programming skills; in addition, we have to attend class twice a week, usually the course only needs to be attended once a week for three and a half hours, this class has been cut in half, although the total number of hours has not changed, the course is much more stressful, the amount of assignments and weekly literature reading has become twice as much as the normal course, for example, after the class on Tuesday, we have to upload the literature reading experience in the course forum by the day before the next class, that is, Wednesday (the second class on the next day, Thursday).

Specific coursework requirements.

Total for the whole semester 38 documents Reading is required (about 5 to 7 per week, concentrated in the first half of the course), reading is required to be uploaded (and with critical thinking, not just combing through the literature for results), and then there are quizzes in class.

Throughout the semester there are 5 door assignment

Diagram visualization tool hands-on (two-person teams on stage report presentation), and

2 hands-on D3 visualizations (one for the basic graph visualization implementation and the other for the interactive visualization of the dashboard implementation), the

literature report( A group of one., Select a piece of literature to present on stage)、

Implement a classical graph visualization algorithm in code (it's the big devil of all assignments, and I spent two full days consumed in the library and still didn't finish it anyway. )

Final project: work in teams of 2 or 3 with industry professionals (get data from them and understand their needs and problems) to develop a complete graph visualization (from drafting to interactive visualization, writing a paper in a standard journal format, presenting it on stage, etc.)

Detailed information and course materials can be viewed and downloaded from this website: https://codydunne.github.io/cs7295-f17

Relational network visualization, in short, is the use of points (nodes) and lines (edges) or matrices (matrices) to present the network relationships between complex data. This series of seven posts (including this one), expected to be posted one a week (begging for prodding hhh), focuses on sorting out what I learned in this class, from reading the literature to actually doing a relational network visualization project at.

Opening Synopsis

brief introduction Relationship Network Visualization Content of the article series。

A general reading of the literature1: node-link( dotted line) & matrix( matrices)

The kinds of relational network visualizations involved can be broadly divided into two types that One uses node-links, the other uses matrices . In the literature readings in the classroom, it is clear to understand that presenting relational network visualizations falls into these two main camps. node-link can present more detailed information, but it tends to complicate the visualization of the relational network. matrix can greatly simplify the relational network and clearly present the relationship between two, but it also loses some important information in the relational network data, such as the inability to present geographical information. The love/hate relationship between the two will be generally sorted out in this one.

a) Node-link dotted line & b) Matrix matrices

Literature Overview #2: Visualizing Visual Representation

Including relational network visualizations, when presenting data visualizations, it is important to note Colors and other visual elementshow to avoid misinterpretation by using the wrong colors, how to use Gestalt rule (Gestalt Principles) to aid visualization, which will be explored in this article.

Using the rainbow color scale (left) is prone to visual misinterpretation

A general reading of the literature3: Presentation of auxiliary perspectives(integration & coordinated views)

This section will describe how other styles of data visualization can be used to complement relational network visualization. Like a dashboard, relational networks will require other visualizations to help users explore and gain insight into the network relationships between data.

remove The visualization of two different views, left and right, can assist the user in understanding the same data

The visualization of two different views, left and right, can assist the user in understanding the same data

Relationship network visualization tool: gephi

Choosing a web visualization tool and reporting on it on stage is one of the assignments for this class. The tools available are : Gephi, NodeXL, Cytoscape, Tulip, Visone, etc. Our group (two people) went with gephi. gephi features in ease of operation and the ability to render dynamic network relationships, etc., but because it is also an open source tool, there are many pitfalls. This post will be a handy guide on how to use the basic features of Gephi and the problems you will encounter.

gephi

literature report

This piece is also one of the assignments for the class, Each person needs to choose one piece of literature( Listed by teacher, Additional literature can be found on the course website atPaper Presentations look over) Read closely and report on stage。 I'll go with《Many-to-Many Geographically-Embedded Flow Visualisation- An Evaluation》, Posted in.2016 year。 For me., This essay is distinguished by the fact that the literature and mathematics cited are quite logical, One can look at the literature review of this article to understand why the authors are proposing a new Relationship Network Visualization Style to solve the problem; additionally, The article explains in detail a new Relationship Network Visualization design process for the style and how to design experiments to test that visualization, And the part where I benefited a lot。

How to trade-off between node-link and matrix is a focus of this literature

End-of-period items

As a key project for the course, I worked with a classmate to make an interactive visualization with a fully structured, already publishable paper from start to finish, from data selection, problem exploration, visualization style brainstorming, and programming use. This post will present the entire project design process.

Screenshot of the end-of-period project

Above, Happy New Year.

Reference.

M. Ghoniem, J.-D.Fekete, and P. Castagliola, “A Comparison of theReadability of Graphs Using Node-Link and Matrix-Based Representations,”IEEE Symposium on Information Visualization, 2004.

Stef van den Elzenand Jarke J. van Wijk, “Multivariate NetworkExploration and Presentation: From Detail to Overview via Selections andAggregations,” 2014.

Michelle A. Borkinet al., “Evaluation of Artery Visualizations for HeartDisease Diagnosis,” 2011.

Yalong Yang et al., “Many-to-ManyGeographically-Embedded Flow Visualisation- An Evaluation,” 2016.

Follow @drawingset

[Great data visualization article recommendation

1. 6 Charts Explaining the State of U.S. Infrastructure

2. [2017 Global University Graduate Job Competitiveness Ranking]

3. Lucky and unlucky numbers in the world

4. Will China's central bank follow suit after the Fed's rate hike?

5. Eight months of disruption in group tours to Korea: Chinese tourists still top source despite halving

6. How big of a house can $500,000 buy in various cities in the United States?

7. Illustration: the fastest flying machine in the universe

8. Graphic: How China's subways will expand rapidly from 1990 to 2020

9. graphic: which movies critics love but audiences don't buy


Recommended>>
1、Chile adopts blockchain technology for national energy grid
2、Breathless Can Zuckerbergs own accounts be hacked and users be spared
3、VernBrownell The best shortcut to facilitate the commercialization of quantum computers
4、The 2017 newborn baby popup names are out Guangdong piles up names Zixuan Zizheng Zixhao
5、Global Goddess Health Guardian Akcome Wins Little Gold

    已推荐到看一看 和朋友分享想法
    最多200字,当前共 发送

    已发送

    朋友将在看一看看到

    确定
    分享你的想法...
    取消

    分享想法到看一看

    确定
    最多200字,当前共

    发送中

    网络异常,请稍后重试

    微信扫一扫
    关注该公众号