最近學(xué)習(xí)Tableau, 感覺(jué)老師的一些可視化觀念很好，現(xiàn)粗淺的翻譯一下，分享給大家：
課程Coursera 杜克大學(xué)Duke University<Data visualization and Communication with Tableau>

Using visualization science to influence business decision

使用可視化科學(xué)影響商務(wù)決策

what people look tight to what people decide
“人們看到的與他們所決定的密切相關(guān)”（百聞不如一見(jiàn)）
guide audience's eyes to where we want them to be
（用可視化）引導(dǎo)觀眾的目光到我們想它們所在的地方
first we need to figure out what story we want to tell
首先我們需要解決我們想要告訴觀眾什么樣的故事

Storyboarding Your Data Story

用故事板講你的數(shù)據(jù)

the storyboarding hourglass

故事板沙漏圖

big picture -- details -- big picture
從藍(lán)圖---細(xì)節(jié)---藍(lán)圖

stortyboarding.png

1. WHY? SO WHAT?
   1.為什么？那又怎樣？
   為什么觀眾要聽(tīng)你的？如果不聽(tīng)的話那又怎么樣？
   motivate audience to listen to what you say is the first important thing you need to do
   首先要做的最重要的是激發(fā)觀眾聽(tīng)你所說(shuō)的的激情
   ("so what" is more important than "what")
   （“那又怎么”比“什么”更重要）
   來(lái)引出你要解決的問(wèn)題

2.Use "SMART" redcommendation to tell audience the detail
3.Summarize the main point you are going to support your recommendation
4.supporting evidence
5.link evidence to SMART rec.

6. What you could gain
7. WHY? SO WHAT?

Make your data story come alive

讓你的故事栩栩如生

story:Characters; Location; Confinct; Resolution
講一個(gè)故事：角色，地點(diǎn)，矛盾，解決

Storyboarding your presentation

用故事板來(lái)做演示

**more companies need storyboard presentation than traditional dashboard **
”越來(lái)越多的公司需要故事板式的演示，而不是傳統(tǒng)的儀表盤(pán)
"post-insight" to compile your storypoints
1.all insights summarize in one sentance
2.figure out three main points and three subpoints
3.organize your story points:if not controversial, most compelling storypoint first
4.draw a quick sketch what kind of visualization for each post-it

Stress-testing your story

給你的故事做壓力測(cè)試

The best stress-testers are teams
get feedback from your data analytics team, and stakeholder

你會(huì)出現(xiàn)的問(wèn)題：

Overgeneralization and sample bias

過(guò)泛化和樣本偏差

we'll overgeneralize because of the lack data, bias data, and big data won't resolve it
tips:1.Always ask questions about collection methods

2. Always check number of data points
3. test subsets of data for consistency(random split subset)
4. look for common characteristics of outlier and missing data

Misinterpretations Due to Lack of controls

由于缺少控制導(dǎo)致的誤解

control group and treatment group, it should be random
avoid:1.overgeneralization and sample bias 2.conclusion without controls

Correlation does not equal causation

相關(guān)性不等于因果性

it's very dangerous for business analysis

How correlations impact business decisions

相關(guān)性是如何影響商務(wù)決策的

if you see correlation between two variables , you should do A/B test
if not test condition:

1. always think if there's a third or fourth variable that explain the relationship you see
2. examine whether the correlation you're basing your business recommendation on exists in other contexts or datasets.
3. come up different complementary angles to assess the relationship
   eg: google flu predict
   if you miss the notice, it will cost you and your company a lot of money.

Tools for Conveying Your Data Story

轉(zhuǎn)化數(shù)據(jù)故事的工具

Choosing visualization for story points

為故事要點(diǎn)選擇可視化

complicated graphs may have risks ,so stick to line charts and bar charts
復(fù)雜的圖片是有風(fēng)險(xiǎn)的，所以專(zhuān)注于線圖和柱圖
when use line chart(change over time or ordered category)
線圖（隨時(shí)間或有順序的類(lèi)別而改變）
when use bar chart (compare categories)
柱圖（比較類(lèi)別）
pie chart(4 or fewer)
餅圖（少于4個(gè)分類(lèi)或兩個(gè)分類(lèi)）
don't use
不要用
scatter plot(for technical audience)
散點(diǎn)圖
3D chart
3D圖

the Neuroscience of visual Perception can make or break your visualization

視覺(jué)捕獲的神經(jīng)學(xué)研究
human are good at measure position and length(that's why us bar and lines)
科學(xué)研究人的視覺(jué)對(duì)高度和位置最為敏感，而對(duì)大小顏色等相對(duì)較弱
因此我們要用柱圖和線圖

Misinterpretations caused by Colorbars

對(duì)顏色圖的使用事項(xiàng)：
don't use color to convey detailed quantitative differences in the values of continuous variables
do use color to illustrate general patterns
code for different categories of categorical variables(to draw attention)
draw attention to something
Visual contrast directs where your audience looks
use color to strength when you want your audience to look

Putting compelling data visualizations into persuasive business presentations

把數(shù)據(jù)視覺(jué)引入商務(wù)演示

Formatting slides to communicate data stories
1.Max data ink ratio(data ink:ink that represent the actual data; Non-data ink:everything else)
let audience focus on the data

2. understanding at a glance
3. fonts
4. make sure everyone in the room can see the slide(when u finish the slide,zoom to 66% to check if can see)
5. Don't make your audience do visual math
6. Use titles to convey take-home messages

Formatting presentations to communicate data stories

轉(zhuǎn)換演示為交流數(shù)據(jù)故事

1. uniform the presentations, but you can different background font ~ in transition slide or summary slide
2. people's attention is 10 mins, so use transitions slide
   the key is to make sure the transitions are smooth, not disruptive
3. animations

1.check for typos!! then check again!
2.bold is more readable that italics or underlining

3. No distorted or fuzzy pictures!
   4.use 2-3 colors.

Delivering your data story

傳遞你的數(shù)據(jù)故事

what you say: the best to call audience is to interact with them.
how you say: voice and tone
what you will do before the presentation: practice,practice,practice.
what you will do during the presentation.
what you will after the presentation: give feedback and receive feedback.