Карта сайта
Версия для печати

Основы создания эфективной визуализации данных

28 ноября 2013 Ежедневно мир создает от 2,5 экзабайтов данных. Поэтому умение визуализировать данные является критичным навыком подачи информации. Графическое отображение данных гарантирует повышенный уровень восприятия информации, ее доступность и понимание на всех уровнях организации, и, следовательно, принятие верных решений в ответ на динамические изменения бизнес-среды. Предлагаем ознакомиться с серией уроков Ноа Ильински, эксперта IBM по визуализации данных. (Материал опубликован на английском языке)
This series will teach you how to design effective visualizations enabling you to understand your data, then identify what's important, and, ultimately, make the right decisions.

Here's a breakdown of each talk:

Talk 1: Four Pillars of Visualization


This talk discusses the broad design considerations necessary for effective visualizations. You will learn what’s required for a visualization to be successful, gain insight for critically evaluating visualizations you encounter, and come away with new ways to think about the visualization design process.

To be most effective, a visualization must have, in this order:
  1. Purpose (Why are we creating this visualization? Who is it for?)
  2. Content (What data matters? What relationships matter?)
  3. Structure (How do we best reveal those data and relationships?)
  4. Formatting (How does it look & feel?)
This talk will define these four pillars, reveal why they must be selected in this order, and discuss the importance and impact each has on your visualization.

(Click here to access Talk 1's slides.)

 
Talk 2: Deeper Visualization Examples

For this talk we'll examine and discuss several examples from the domains of time series data, geographic data, and network data. Some of the examples show common approaches, while most are different and interesting takes on how to represent these common data types in uncommon ways. The goal of this talk is to provide ideas, and to show that there are alternate structures available that will highlight different aspects of data.

(Click here to access Talk 2's slides.)


Talk 3: Choosing the Right Structure for Your Visualization
There are many ways to visualize any given data set. Choosing the right structure makes the difference between highlighting the right data and hiding it. In this talk we'll review common, archetypal visualization structures, discuss the strengths of each and when each is appropriate, and talk about some structures you should never use.

(Click here to access Talk 3's slides.)

 
Talk 4: Choosing Good Encodings for Your Visualization

Date: Tuesday, December 3, 2013

Time: 1pm Eastern Time (New York City, USA


Should color represent region or rate? Would you use size to represent flavor or sweetness? Can position represent both categories and quantities at the same time? There are (usually) right answers for all of these questions, based in the science of how we perceive various visual properties. This talk reveals the foundations of good encoding, and dives deeply into the encoding cheat-sheet discussed in lecture #1.

 
Talk 5: Critique

Bring your or a coworker’s visualization to this session. In a constructive way, we’ll apply the principles we learned in our previous sessions to improve the effectiveness of your visualization or extol its virtues as an excellent example of an effective visualization.